Object Detection Evaluation 2012


The object detection and object orientation estimation benchmark consists of 7481 training images and 7518 test images, comprising a total of 80.256 labeled objects. All images are color and saved as png. For evaluation, we compute precision-recall curves for object detection and orientation-similarity-recall curves for joint object detection and orientation estimation. In the latter case not only the object 2D bounding box has to be located correctly, but also the orientation estimate in bird's eye view is evaluated. To rank the methods we compute average precision and average orientation similiarity. We require that all methods use the same parameter set for all test pairs. Our development kit provides details about the data format as well as MATLAB / C++ utility functions for reading and writing the label files.

We evaluate object detection performance using the PASCAL criteria and object detection and orientation estimation performance using the measure discussed in our CVPR 2012 publication. For cars we require an overlap of 70%, while for pedestrians and cyclists we require an overlap of 50% for a detection. Detections in don't care areas or detections which are smaller than the minimum size do not count as false positive. Difficulties are defined as follows:

  • Easy: Min. bounding box height: 40 Px, Max. occlusion level: Fully visible, Max. truncation: 15 %
  • Moderate: Min. bounding box height: 25 Px, Max. occlusion level: Partly occluded, Max. truncation: 30 %
  • Hard: Min. bounding box height: 25 Px, Max. occlusion level: Difficult to see, Max. truncation: 50 %

All methods are ranked based on the moderately difficult results. Note that for the hard evaluation ~2 % of the provided bounding boxes have not been recognized by humans, thereby upper bounding recall at 98 %. Hence, the hard evaluation is only given for reference.
Note 1: On 25.04.2017, we have fixed a bug in the object detection evaluation script. As of now, the submitted detections are filtered based on the min. bounding box height for the respective category which we have been done before only for the ground truth detections, thus leading to false positives for the category "Easy" when bounding boxes of height 25-39 Px were submitted (and to false positives for all categories if bounding boxes smaller than 25 Px were submitted). We like to thank Amy Wu, Matt Wilder, Pekka Jänis and Philippe Vandermersch for their feedback. The last leaderboards right before the changes can be found here!

Additional information used by the methods
  • Stereo: Method uses left and right (stereo) images
  • Flow: Method uses optical flow (2 temporally adjacent images)
  • Multiview: Method uses more than 2 temporally adjacent images
  • Laser Points: Method uses point clouds from Velodyne laser scanner
  • Additional training data: Use of additional data sources for training (see details)

Car


Method Setting Code Moderate Easy Hard Runtime Environment
1 iDST-VC 90.55 % 90.88 % 81.04 % 4 s GPU @ 2.5 Ghz (Python + C/C++)
2 BM-NET 90.48 % 90.83 % 80.63 % 4.0 s GPU @ 2.5 Ghz (C/C++)
3 SAITv1 90.36 % 90.78 % 80.48 % 0.18 s GPU @ 2.5 Ghz (Python, C/C++)
4 TuSimple code 90.33 % 90.77 % 82.86 % 1.6 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit all the layers: Fast and accurate cnn object detector with scale dependent pooling and cascaded rejection classifiers. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
K. He, X. Zhang, S. Ren and J. Sun: Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition 2016.
5 THU CV-AI 90.31 % 90.75 % 72.20 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
6 RRC code 90.22 % 90.61 % 87.44 % 3.6 s GPU @ 2.5 Ghz (Python + C/C++)
J. Ren, X. Chen, J. Liu, W. Sun, J. Pang, Q. Yan, Y. Tai and L. Xu: Accurate Single Stage Detector Using Recurrent Rolling Convolution. CVPR 2017.
7 cfenet 90.08 % 90.30 % 84.70 % 4.0 s GPU @ 2.5 Ghz (Python + C/C++)
8 SJTU-HW 90.08 % 90.81 % 79.98 % 0.85 s GPU @ 1.5 Ghz (Python + C/C++)
9 SWC 90.05 % 90.82 % 80.59 % 0.5 s GPU @ >3.5 Ghz (Python + C/C++)
10 Deep MANTA 90.03 % 97.25 % 80.62 % 0.7 s GPU @ 2.5 Ghz (Python + C/C++)
F. Chabot, M. Chaouch, J. Rabarisoa, C. Teulière and T. Chateau: Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image. CVPR 2017.
11 lpm 90.03 % 90.75 % 80.99 % 1 s 4 cores @ 3.5 Ghz (C/C++)
12 sensekitti code 90.00 % 90.76 % 81.83 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
13 F-PointNet
This method makes use of Velodyne laser scans.
code 90.00 % 90.78 % 80.80 % 0.17 s GPU @ 3.0 Ghz (Python)
C. Qi, W. Liu, C. Wu, H. Su and L. Guibas: Frustum PointNets for 3D Object Detection from RGB-D Data. arXiv preprint arXiv:1711.08488 2017.
14 SAITv2 89.91 % 95.04 % 79.91 % 0.07 s GPU @ 2.5 Ghz (Python + C/C++)
15 SG 89.89 % 90.51 % 80.67 % 0.04 s 1 core @ 2.5 Ghz (C/C++)
16 M3D 89.88 % 90.59 % 80.61 % 0.4 s GPU @ 2.5 Ghz (Python + C/C++)
17 MBR-SSD 89.82 % 90.32 % 82.28 % 4.0 s GPU @ 2.5 Ghz (Python + C/C++)
18 CNN 89.81 % 90.50 % 80.64 % 1 s 1 core @ 2.5 Ghz (C/C++)
19 SINet+ code 89.73 % 90.51 % 77.82 % 0.3 s TITAN X GPU
X. Hu, X. Xu, Y. Xiao, H. Chen, S. He, J. Qin and P. Heng: SINet: A Scale-insensitive Convolutional Neural Network for Fast Vehicle Detection. arXiv preprint arXiv:1804.00433 2018.
20 Aston-EAS 89.64 % 90.49 % 77.95 % 0.24 s 8 cores @ >3.5 Ghz (Python + C/C++)
21 Paul-Fr-RCNN 89.59 % 90.76 % 77.23 % 0.12 s 1 core @ 2.5 Ghz (C/C++)
22 D3D 89.59 % 90.51 % 80.57 % 0.4 s 1 core @ 3.5 Ghz (Python)
23 SINet_VGG code 89.56 % 90.60 % 78.19 % 0.2 s TITAN X GPU
X. Hu, X. Xu, Y. Xiao, H. Chen, S. He, J. Qin and P. Heng: SINet: A Scale-insensitive Convolutional Neural Network for Fast Vehicle Detection. arXiv preprint arXiv:1804.00433 2018.
24 DF-PC_CNN
This method makes use of Velodyne laser scans.
89.45 % 90.78 % 82.77 % 0.5 s GPU @ 3.0 Ghz (Matlab + C/C++)
25 SDP+RPN 89.42 % 89.90 % 78.54 % 0.4 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in Neural Information Processing Systems 2015.
26 RaC 89.39 % 90.02 % 80.29 % 1s s GPU @ 1.0 Ghz (C/C++)
27 VCTNet 89.23 % 89.82 % 79.92 % 0.18 s GPU @ 3.5 GHz (C/C++)
28 ITVD 89.23 % 90.57 % 79.31 % 0.3 s GPU @ 2.5 Ghz (C/C++)
29 Sogo_MM 89.17 % 90.80 % 79.58 % 1.5 s GPU @ 2.5 Ghz (C/C++)
30 MV3D
This method makes use of Velodyne laser scans.
89.17 % 90.53 % 80.16 % 0.36 s GPU @ 2.5 Ghz (Python + C/C++)
X. Chen, H. Ma, J. Wan, B. Li and T. Xia: Multi-View 3D Object Detection Network for Autonomous Driving. CVPR 2017.
31 VAT-Net 89.11 % 90.53 % 79.63 % 1 s GPU @ 2.5 Ghz (Python + C/C++)
32 SINet_PVA code 89.08 % 90.44 % 75.85 % 0.11 s TITAN X GPU
X. Hu, X. Xu, Y. Xiao, H. Chen, S. He, J. Qin and P. Heng: SINet: A Scale-insensitive Convolutional Neural Network for Fast Vehicle Detection. arXiv preprint arXiv:1804.00433 2018.
33 LTT 89.00 % 90.16 % 81.94 % 0.4 s 1 core @ 3.5 Ghz (Python)
34 HSR2 88.98 % 90.76 % 78.62 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
35 R-DML 88.92 % 90.42 % 79.57 % 0.16 s GPU @ 2.5 Ghz (Python + C/C++)
36 SubCNN 88.86 % 90.75 % 79.24 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
37 Deep3DBox 88.86 % 90.47 % 77.60 % 1.5 s GPU @ 2.5 Ghz (C/C++)
A. Mousavian, D. Anguelov, J. Flynn and J. Kosecka: 3D Bounding Box Estimation Using Deep Learning and Geometry. CVPR 2017.
38 desNet 88.85 % 90.51 % 79.36 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
39 FMLA 88.83 % 90.45 % 77.04 % 0.17 s GPU @ 1.5 Ghz (C/C++)
40 MS-CNN code 88.83 % 90.46 % 74.76 % 0.4 s GPU @ 2.5 Ghz (C/C++)
Z. Cai, Q. Fan, R. Feris and N. Vasconcelos: A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection. ECCV 2016.
41 vfnet 88.77 % 89.63 % 79.55 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
42 DeepStereoOP 88.75 % 90.34 % 79.39 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
43 SN-net 88.60 % 88.97 % 79.24 % 0.02 s GPU @ 2.5 Ghz (Python + C/C++)
44 eCHIEV 88.55 % 90.42 % 79.79 % 1.5 s 1 core @ 2.5 Ghz (C/C++)
45 DesNet 88.47 % 89.49 % 79.09 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
46 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
88.40 % 90.40 % 80.21 % 0.06 s GPU @ 1.8 Ghz (Python)
47 RCNN 88.36 % 89.74 % 72.64 % 1 s 1 core @ 2.5 Ghz (C/C++)
48 3DOP
This method uses stereo information.
code 88.34 % 90.09 % 78.79 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
49 HM3D 88.26 % 89.86 % 78.24 % 0.35 s GPU @ >3.5 Ghz (C/C++)
50 MM-MRFC
This method uses optical flow information.
This method makes use of Velodyne laser scans.
88.20 % 90.93 % 78.02 % 0.05 s GPU @ 2.5 Ghz (C/C++)
A. Costea, R. Varga and S. Nedevschi: Fast Boosting based Detection using Scale Invariant Multimodal Multiresolution Filtered Features. CVPR 2017.
51 AVOD
This method makes use of Velodyne laser scans.
code 88.08 % 89.73 % 80.14 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
52 InNet 88.00 % 87.60 % 78.85 % 0.16 s GPU @ 3.5 Ghz (Python + C/C++)
53 Mono3D code 87.86 % 90.27 % 78.09 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
54 WRInception 87.62 % 88.98 % 77.52 % 0.06 s GPU @ 2.5 Ghz (C/C++)
55 AVOD-FPN
This method makes use of Velodyne laser scans.
code 87.44 % 89.99 % 80.05 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
56 denet 87.35 % 87.56 % 78.32 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
57 MonoFusion 87.33 % 90.43 % 76.78 % 0.12 s TITAN X GPU
58 CNN-ds code 87.15 % 86.86 % 70.77 % 0.05 s 1 core @ 2.5 Ghz (Python)
59 RCL-FC 86.56 % 90.25 % 71.26 % 0.19 s GPU @ 2.5 Ghz (Matlab + C/C++)
60 VxNet(LiDAR)
This method makes use of Velodyne laser scans.
85.95 % 90.30 % 79.21 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
61 ReSqueeze 85.74 % 87.12 % 77.02 % 0.03 s GPU @ >3.5 Ghz (Python)
62 AVOD-SSD
This method makes use of Velodyne laser scans.
code 85.71 % 88.94 % 78.05 % 0.09 s GPU @ 2.5 Ghz (Python)
63 YOLOv2-3cls 85.65 % 88.01 % 74.16 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
64 anm 85.33 % 90.11 % 76.55 % 3 s 1 core @ 2.5 Ghz (C/C++)
65 MMOD+CNN code 83.14 % 89.86 % 69.29 % 0.28 s 4 cores @ >3.5 Ghz (C/C++)
66 R-RRC 82.94 % 89.99 % 72.21 % 0.09 s GPU @ 1.0 Ghz (Python + C/C++)
67 LPN 81.67 % 87.70 % 72.69 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
68 A3DODWTDA (image) code 81.54 % 76.21 % 66.85 % 0.8 s GPU @ 3.0 Ghz (Python)
F. Gustafsson and E. Linder-Norén: Automotive 3D Object Detection Without Target Domain Annotations. 2018.
69 ISSD 81.39 % 88.08 % 72.94 % 0.045s GPU @ 3.0 Ghz (Python + C/C++)
70 SDP+CRC (ft) 81.33 % 90.39 % 70.33 % 0.6 s GPU @ 2.5 Ghz (C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
71 deprecated 81.29 % 85.23 % 69.32 % 0.00 s GPU @ 2.5 Ghz (C/C++)
72 Fast R-RRC 80.80 % 89.23 % 71.47 % 0.058 s GPU @ 1.0 Ghz (Python + C/C++)
73 vf-ssd 80.35 % 75.36 % 73.76 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
74 MV3D (LIDAR)
This method makes use of Velodyne laser scans.
79.76 % 89.80 % 78.61 % 0.24 s GPU @ 2.5 Ghz (Python + C/C++)
X. Chen, H. Ma, J. Wan, B. Li and T. Xia: Multi-View 3D Object Detection Network for Autonomous Driving. CVPR 2017.
75 FPN 79.48 % 89.45 % 69.81 % 5 s 1 core @ 2.5 Ghz (C/C++)
76 RFCN 79.44 % 88.69 % 70.06 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
77 Denet 79.30 % 88.42 % 69.92 % 0.06 s 1 core @ 2.5 Ghz (C/C++)
78 rtd 79.23 % 87.81 % 69.52 % 0.01 s 1 core @ 2.5 Ghz (Python)
79 RefineNet 79.21 % 90.16 % 65.71 % 0.20 s GPU @ 2.5 Ghz (Matlab + C++)
R. Rajaram, E. Bar and M. Trivedi: RefineNet: Iterative Refinement for Accurate Object Localization. Intelligent Transportation Systems Conference 2016.
80 Faster R-CNN code 79.11 % 87.90 % 70.19 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards Real- Time Object Detection with Region Proposal Networks. NIPS 2015.
81 Kiwoo 79.06 % 89.23 % 70.50 % 0.1 s 1 core @ 2.5 Ghz (Python + C/C++)
82 FRCNN+Or code 78.95 % 89.87 % 68.97 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
83 fd3 78.91 % 86.73 % 70.29 % 0.01 s GPU @ 2.5 Ghz (C/C++)
84 T2Method 78.26 % 88.55 % 69.76 % 0.05 s GPU @ 2.5 Ghz (Python + C/C++)
85 MB-Net 77.91 % 86.31 % 61.22 % 0.02 s GPU @ 1.5 Ghz (C/C++)
86 HM 77.72 % 87.90 % 61.36 % 1 s 1 core @ 2.5 Ghz (C/C++)
87 avodC 77.54 % 86.86 % 70.00 % 0.1 s GPU @ 2.5 Ghz (Python)
88 spLBP 77.39 % 80.16 % 60.59 % 1.5 s 8 cores @ 2.5 Ghz (Matlab + C/C++)
Q. Hu, S. Paisitkriangkrai, C. Shen, A. Hengel and F. Porikli: Fast Detection of Multiple Objects in Traffic Scenes With a Common Detection Framework. IEEE Trans. Intelligent Transportation Systems 2016.
89 SceneNet 77.34 % 87.90 % 68.38 % 0.03 s GPU @ 2.5 Ghz (C/C++)
90 MTDP 76.91 % 84.24 % 67.91 % 0.15 s GPU @ 2.0 Ghz (Python)
91 NLK 76.81 % 85.26 % 74.78 % 0.05 s 1 core @ 2.5 Ghz (Python + C/C++)
92 FYSqueeze 76.73 % 84.06 % 67.96 % 0.01 s >8 cores @ 2.5 Ghz (Python)
93 Reinspect code 76.65 % 88.36 % 66.56 % 2s 1 core @ 2.5 Ghz (C/C++)
R. Stewart, M. Andriluka and A. Ng: End-to-End People Detection in Crowded Scenes. CVPR 2016.
94 Regionlets 76.56 % 86.50 % 59.82 % 1 s >8 cores @ 2.5 Ghz (C/C++)
X. Wang, M. Yang, S. Zhu and Y. Lin: Regionlets for Generic Object Detection. T-PAMI 2015.
W. Zou, X. Wang, M. Sun and Y. Lin: Generic Object Detection with Dense Neural Patterns and Regionlets. British Machine Vision Conference 2014.
C. Long, X. Wang, G. Hua, M. Yang and Y. Lin: Accurate Object Detection with Location Relaxation and Regionlets Relocalization. Asian Conference on Computer Vision 2014.
95 Inha_cvlab 76.04 % 84.38 % 67.32 % 0.01 s GPU @ 2.5 Ghz (Python)
96 AOG code 75.97 % 85.58 % 60.96 % 3 s 4 cores @ 2.5 Ghz (Matlab)
T. Wu, B. Li and S. Zhu: Learning And-Or Models to Represent Context and Occlusion for Car Detection and Viewpoint Estimation. TPAMI 2016.
B. Li, T. Wu and S. Zhu: Integrating Context and Occlusion for Car Detection by Hierarchical And-Or Model. ECCV 2014.
97 VS3D 75.84 % 83.92 % 60.24 % 0.58 s GPU @ 2.5 Ghz (C/C++)
98 3D FCN
This method makes use of Velodyne laser scans.
75.83 % 85.54 % 68.30 % >5 s 1 core @ 2.5 Ghz (C/C++)
B. Li: 3D Fully Convolutional Network for Vehicle Detection in Point Cloud. IROS 2017.
99 3D-SSMFCNN code 75.78 % 75.51 % 67.75 % 0.1 s GPU @ 1.5 Ghz (C/C++)
L. Novak: Vehicle Detection and Pose Estimation for Autonomous Driving. 2017.
100 3DVP code 75.77 % 81.46 % 65.38 % 40 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Data-Driven 3D Voxel Patterns for Object Category Recognition. IEEE Conference on Computer Vision and Pattern Recognition 2015.
101 Pose-RCNN 75.74 % 88.89 % 61.86 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
102 AR-FCN 75.49 % 81.24 % 66.00 % 0.19 s GPU @ 2.5 Ghz (C/C++)
103 SubCat code 75.46 % 81.45 % 59.71 % 0.7 s 6 cores @ 3.5 Ghz (Matlab + C/C++)
E. Ohn-Bar and M. Trivedi: Learning to Detect Vehicles by Clustering Appearance Patterns. T-ITS 2015.
104 DimStr-LKY 75.22 % 81.21 % 67.28 % 0.1 s GPU @ 2.5 Ghz (Matlab + C/C++)
105 Roadstar.ai 74.84 % 82.93 % 67.18 % 0.08 s GPU @ 2.0 Ghz (Python)
106 A3DODWTDA
This method makes use of Velodyne laser scans.
code 74.71 % 78.21 % 66.70 % 0.08 s GPU @ 3.0 Ghz (Python)
F. Gustafsson and E. Linder-Norén: Automotive 3D Object Detection Without Target Domain Annotations. 2018.
107 FD2 74.68 % 87.14 % 65.70 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
108 3dSSD 74.53 % 83.54 % 67.59 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
109 CHTTL 73.54 % 80.68 % 65.43 % 0.07 s 1 core @ 2.5 Ghz (Python)
110 tiny-det 73.46 % 81.88 % 63.70 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
111 3DVSSD 73.39 % 84.39 % 65.64 % 0.06 s 1 core @ 2.5 Ghz (C/C++)
112 FD 72.64 % 82.34 % 60.31 % 0.01 s GPU @ >3.5 Ghz (Python)
113 MV-RGBD-RF
This method makes use of Velodyne laser scans.
69.92 % 76.49 % 57.47 % 4 s 4 cores @ 2.5 Ghz (C/C++)
A. Gonzalez, D. Vazquez, A. Lopez and J. Amores: On-Board Object Detection: Multicue, Multimodal, and Multiview Random Forest of Local Experts.. IEEE Trans. on Cybernetics 2016.
A. Gonzalez, G. Villalonga, J. Xu, D. Vazquez, J. Amores and A. Lopez: Multiview Random Forest of Local Experts Combining RGB and LIDAR data for Pedestrian Detection. IEEE Intelligent Vehicles Symposium (IV) 2015.
114 AOG-View 69.89 % 84.29 % 57.25 % 3 s 1 core @ 2.5 Ghz (Matlab, C/C++)
B. Li, T. Wu and S. Zhu: Integrating Context and Occlusion for Car Detection by Hierarchical And-Or Model. ECCV 2014.
115 YOLOv2 code 69.01 % 86.40 % 59.57 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
116 tester 68.85 % 78.94 % 62.32 % 0.1
117 Vote3Deep
This method makes use of Velodyne laser scans.
68.39 % 76.95 % 63.22 % 1.5 s 4 cores @ 2.5 Ghz (C/C++)
M. Engelcke, D. Rao, D. Zeng Wang, C. Hay Tong and I. Posner: Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks. ArXiv e-prints 2016.
118 GPVL 67.89 % 77.76 % 58.23 % 10 s 1 core @ 2.5 Ghz (C/C++)
119 BdCost48LDCF code 67.08 % 77.93 % 51.15 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
120 OC-DPM 66.45 % 76.16 % 53.70 % 10 s 8 cores @ 2.5 Ghz (Matlab)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Occlusion Patterns for Object Class Detection. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2013.
121 DPM-VOC+VP 66.25 % 80.45 % 49.86 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
122 HNet code 66.00 % 77.09 % 53.89 % 1 s 1 core @ 2.5 Ghz (Python + C/C++)
123 BdCost48-25C 65.95 % 78.21 % 51.23 % 4 s 1 core @ 2.5 Ghz (C/C++)
124 bin 64.39 % 77.58 % 56.33 % 15ms s GPU @ >3.5 Ghz (Python)
125 MDPM-un-BB 64.20 % 77.32 % 50.18 % 60 s 4 core @ 2.5 Ghz (MATLAB)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
126 BNet 63.24 % 75.09 % 56.05 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
127 PDV-Subcat 63.15 % 77.33 % 49.75 % 7 s 1 core @ 2.5 Ghz (C/C++)
J. Shen, X. Zuo, J. Li, W. Yang and H. Ling: A novel pixel neighborhood differential statistic feature for pedestrian and face detection . Pattern Recognition 2017.
128 GNN 62.59 % 76.03 % 50.18 % 0.2 s 1 core @ 2.5 Ghz (Python)
129 NMRDO 61.72 % 79.48 % 54.06 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
130 YOLOv2 code 61.31 % 76.79 % 50.25 % 0.03 s TITAN X GPU
J. Redmon and A. Farhadi: YOLO9000: better, faster, stronger. arXiv preprint 2016.
131 DPM-C8B1
This method uses stereo information.
60.99 % 74.95 % 47.16 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
132 SubCat48LDCF code 60.53 % 78.16 % 43.66 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
133 Fast-SSD 60.24 % 83.39 % 51.96 % 0.06 s GTX650Ti
134 SAMME48LDCF code 58.50 % 76.22 % 47.50 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
135 GNN 58.29 % 76.26 % 49.96 % 0.2 s 1 core @ 2.5 Ghz (Python)
136 BirdNet
This method makes use of Velodyne laser scans.
57.47 % 78.18 % 56.66 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
137 LSVM-MDPM-sv 57.44 % 71.70 % 46.58 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
138 Faster RCNN 56.58 % 62.31 % 45.27 % 0.11 s GPU @ 2.5 Ghz (Python)
139 LSVM-MDPM-us code 56.10 % 70.52 % 42.87 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
140 ACF-SC 55.76 % 69.76 % 46.27 % <0.3 s 1 core @ >3.5 Ghz (Matlab + C/C++)
C. Cadena, A. Dick and I. Reid: A Fast, Modular Scene Understanding System using Context-Aware Object Detection. Robotics and Automation (ICRA), 2015 IEEE International Conference on 2015.
141 FRO 53.78 % 70.96 % 46.00 % 0.19 s GPU @ 2.5 Ghz (Python)
142 VeloFCN
This method makes use of Velodyne laser scans.
53.45 % 70.68 % 46.90 % 1 s GPU @ 2.5 Ghz (Python + C/C++)
B. Li, T. Zhang and T. Xia: Vehicle Detection from 3D Lidar Using Fully Convolutional Network. RSS 2016 .
143 ACF 52.81 % 62.82 % 43.89 % 0.2 s 1 core @ >3.5 Ghz (Matlab + C/C++)
P. Doll\'ar, R. Appel, S. Belongie and P. Perona: Fast Feature Pyramids for Object Detection. PAMI 2014.
P. Doll\'ar: Piotr's Image and Video Matlab Toolbox (PMT). .
144 SDN
This method makes use of Velodyne laser scans.
52.03 % 71.75 % 47.08 % 0.096 s GPU @ 1.7 Ghz (Python)
145 F-PC_CNN
This method makes use of Velodyne laser scans.
48.61 % 65.73 % 47.67 % 0.5 s GPU @ 3.0 Ghz (Matlab + C/C++)
X. Du, M. Jr., S. Karaman and D. Rus: A General Pipeline for 3D Detection of Vehicles. 2018.
146 Vote3D
This method makes use of Velodyne laser scans.
48.05 % 56.66 % 42.64 % 0.5 s 4 cores @ 2.8 Ghz (C/C++)
D. Wang and I. Posner: Voting for Voting in Online Point Cloud Object Detection. Proceedings of Robotics: Science and Systems 2015.
147 Multimodal Detection
This method makes use of Velodyne laser scans.
code 46.77 % 64.04 % 39.38 % 0.06 s GPU @ 3.5 Ghz (Matlab + C/C++)
A. Asvadi, L. Garrote, C. Premebida, P. Peixoto and U. Nunes: Multimodal vehicle detection: fusing 3D- LIDAR and color camera data. Pattern Recognition Letters 2017.
148 LMNetV2
This method makes use of Velodyne laser scans.
44.20 % 59.58 % 37.90 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
149 VoxelNet basic
This method makes use of Velodyne laser scans.
43.44 % 45.07 % 39.59 % 0.07 s GPU (Python)
150 RT3D
This method makes use of Velodyne laser scans.
39.71 % 49.96 % 41.47 % 0.09 s GPU @ 1.8Ghz
151 LMnetV1.1 36.88 % 53.16 % 30.47 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
152 LMnet
This method makes use of Velodyne laser scans.
36.06 % 51.10 % 30.09 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
153 YOLO 35.86 % 49.47 % 29.74 % 0.03 s GPU @ 1.0 Ghz (C/C++)
154 Licar
This method makes use of Velodyne laser scans.
33.89 % 41.60 % 35.17 % 0.09 s GPU @ 2.0 Ghz (Python)
155 DoBEM 33.61 % 36.35 % 37.78 % 0.6 s GPU @ 2.5 Ghz (Python + C/C++)
S. Yu, T. Westfechtel, R. Hamada, K. Ohno and S. Tadokoro: Vehicle Detection and Localization on Bird's Eye View Elevation Images Using Convolutional Neural Network. IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR) 2017.
156 Kyolo3 33.01 % 47.18 % 27.57 % 0.16 s 4 cores @ 2.5 Ghz (Python)
157 fastRand code 27.83 % 35.24 % 22.33 % 0.05 s 1 core @ 2.5 Ghz (Matlab + C/C++)
158 CSoR
This method makes use of Velodyne laser scans.
code 26.13 % 35.24 % 22.69 % 3.5 s 4 cores @ >3.5 Ghz (Python + C/C++)
L. Plotkin: PyDriver: Entwicklung eines Frameworks für räumliche Detektion und Klassifikation von Objekten in Fahrzeugumgebung. 2015.
159 R-CNN_VGG 26.04 % 32.23 % 20.93 % 10 s GPU @ 2.5 Ghz (Matlab + C/C++)
160 FCN-Depth code 25.66 % 50.55 % 24.95 % 1 s GPU @ 1.5 Ghz (Matlab + C/C++)
161 mBoW
This method makes use of Velodyne laser scans.
23.76 % 37.63 % 18.44 % 10 s 1 core @ 2.5 Ghz (C/C++)
J. Behley, V. Steinhage and A. Cremers: Laser-based Segment Classification Using a Mixture of Bag-of-Words. Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2013.
162 DepthCN
This method makes use of Velodyne laser scans.
code 23.21 % 37.59 % 18.00 % 2.3 s GPU @ 3.5 Ghz (Matlab)
A. Asvadi, L. Garrote, C. Premebida, P. Peixoto and U. Nunes: DepthCN: vehicle detection using 3D- LIDAR and convnet. IEEE ITSC 2017.
163 YOLOv2 code 19.31 % 28.37 % 15.94 % 0.02 s GPU @ 3.5 Ghz (C/C++)
J. Redmon, S. Divvala, R. Girshick and A. Farhadi: You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
164 SPC
This method makes use of Velodyne laser scans.
18.83 % 25.30 % 17.29 % 0.4 s 4 cores @ 2.5 Ghz (Python)
165 TopNet-HighRes
This method makes use of Velodyne laser scans.
15.72 % 18.65 % 15.03 % 0.27 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
166 TopNet-DecayRate
This method makes use of Velodyne laser scans.
14.53 % 17.50 % 14.17 % 0.12 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
167 LidarNet
This method makes use of Velodyne laser scans.
2.66 % 1.98 % 2.18 % 0.007 s GPU @ 2.5 Ghz (C/C++)
Table as LaTeX | Only published Methods

Pedestrian


Method Setting Code Moderate Easy Hard Runtime Environment
1 iDST-VC 80.90 % 88.13 % 74.08 % 1.5 s GPU @ 2.5 Ghz (Python + C/C++)
2 SWC 78.65 % 87.06 % 73.92 % 0.5 s GPU @ >3.5 Ghz (Python + C/C++)
3 F-PointNet
This method makes use of Velodyne laser scans.
code 77.25 % 87.81 % 74.46 % 0.17 s GPU @ 3.0 Ghz (Python)
C. Qi, W. Liu, C. Wu, H. Su and L. Guibas: Frustum PointNets for 3D Object Detection from RGB-D Data. arXiv preprint arXiv:1711.08488 2017.
4 TuSimple code 77.04 % 86.78 % 72.40 % 1.6 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit all the layers: Fast and accurate cnn object detector with scale dependent pooling and cascaded rejection classifiers. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
K. He, X. Zhang, S. Ren and J. Sun: Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition 2016.
5 VCTNet 75.88 % 84.03 % 71.60 % 0.18 s GPU @ 3.5 GHz (C/C++)
6 Argus_detection_v1 75.51 % 83.49 % 71.24 % 0.25 s GPU @ 1.5 Ghz (C/C++)
7 SiRtAKI 75.47 % 86.54 % 68.27 % 0.18 s GPU @ >3.5 Ghz (C/C++)
8 RRC code 75.33 % 84.14 % 70.39 % 3.6 s GPU @ 2.5 Ghz (Python + C/C++)
J. Ren, X. Chen, J. Liu, W. Sun, J. Pang, Q. Yan, Y. Tai and L. Xu: Accurate Single Stage Detector Using Recurrent Rolling Convolution. CVPR 2017.
9 iFDT 74.83 % 86.02 % 70.55 % 2.4 s GPU @ 2.5 Ghz (Python + C/C++)
10 Aston-EAS 74.52 % 85.12 % 69.35 % 0.24 s 8 cores @ >3.5 Ghz (Python + C/C++)
11 SJTU-HW 74.24 % 85.42 % 69.34 % 0.85 s GPU @ 1.5 Ghz (Python + C/C++)
12 FMLA 73.75 % 83.86 % 68.06 % 0.17 s GPU @ 1.5 Ghz (C/C++)
13 MS-CNN code 73.62 % 83.70 % 68.28 % 0.4 s GPU @ 2.5 Ghz (C/C++)
Z. Cai, Q. Fan, R. Feris and N. Vasconcelos: A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection. ECCV 2016.
14 LFF 73.04 % 82.91 % 67.77 % 1 s GPU
15 RCL-FC 72.78 % 82.75 % 67.53 % 0.19 s GPU @ 2.5 Ghz (Matlab + C/C++)
16 eCHIEV 72.71 % 84.27 % 66.65 % 1.5 s 1 core @ 2.5 Ghz (C/C++)
17 FRCNN 72.64 % 82.52 % 69.21 % 1 s >8 cores @ 2.5 Ghz (Python + C/C++)
18 Sogo_MM 71.84 % 83.45 % 67.00 % 1.5 s GPU @ 2.5 Ghz (C/C++)
19 GN 71.55 % 80.73 % 64.82 % 1 s GPU @ 2.5 Ghz (Matlab + C/C++)
S. Jung and K. Hong: Deep network aided by guiding network for pedestrian detection. Pattern Recognition Letters 2017.
20 SubCNN 71.34 % 83.17 % 66.36 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
21 IVA code 70.63 % 83.03 % 64.68 % 0.4 s GPU @ 2.5 Ghz (C/C++)
Y. Zhu, J. Wang, C. Zhao, H. Guo and H. Lu: Scale-adaptive Deconvolutional Regression Network for Pedestrian Detection. ACCV 2016.
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in neural information processing systems 2015.
22 SDP+RPN 70.20 % 79.98 % 64.84 % 0.4 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in Neural Information Processing Systems 2015.
23 MM-MRFC
This method uses optical flow information.
This method makes use of Velodyne laser scans.
69.96 % 82.37 % 64.76 % 0.05 s GPU @ 2.5 Ghz (C/C++)
A. Costea, R. Varga and S. Nedevschi: Fast Boosting based Detection using Scale Invariant Multimodal Multiresolution Filtered Features. CVPR 2017.
24 WRInception 68.76 % 79.98 % 63.48 % 0.06 s GPU @ 2.5 Ghz (C/C++)
25 3DOP
This method uses stereo information.
code 67.46 % 82.36 % 64.71 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
26 DeepStereoOP 67.32 % 82.50 % 65.14 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
27 sensekitti code 67.28 % 80.12 % 62.25 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
28 VAT-Net 67.15 % 77.28 % 59.59 % 0.08 s GPU @ 2.5 Ghz (Python)
29 HNet code 66.74 % 77.39 % 62.26 % 1 s 1 core @ 2.5 Ghz (Python + C/C++)
30 Mono3D code 66.66 % 77.30 % 63.44 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
31 HM3D 65.97 % 77.60 % 61.09 % 0.35 s GPU @ >3.5 Ghz (C/C++)
32 HSR2 65.91 % 78.05 % 63.05 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
33 Faster R-CNN code 65.91 % 78.35 % 61.19 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards Real- Time Object Detection with Region Proposal Networks. NIPS 2015.
34 R-DML 64.82 % 77.15 % 60.76 % 0.16 s GPU @ 2.5 Ghz (Python + C/C++)
35 SDP+CRC (ft) 64.25 % 77.81 % 59.31 % 0.6 s GPU @ 2.5 Ghz (C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
36 PCN 63.48 % 77.88 % 58.59 % 0.6 s
37 Pose-RCNN 63.38 % 77.69 % 57.42 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
38 CFM 63.26 % 74.21 % 56.44 % <2 s GPU @ 2.5 Ghz (Matlab + C/C++)
Q. Hu, P. Wang, C. Shen, A. Hengel and F. Porikli: Pushing the Limits of Deep CNNs for Pedestrian Detection. IEEE Transactions on Circuits and Systems for Video Technology 2017.
39 RPN+BF code 61.29 % 75.58 % 56.08 % 0.6 s GPU @ 2.5 Ghz (Matlab + C/C++)
L. Zhang, L. Lin, X. Liang and K. He: Is Faster R-CNN Doing Well for Pedestrian Detection?. ECCV 2016.
40 ReSqueeze 61.25 % 72.78 % 57.43 % 0.03 s GPU @ >3.5 Ghz (Python)
41 Regionlets 61.16 % 72.96 % 55.22 % 1 s >8 cores @ 2.5 Ghz (C/C++)
X. Wang, M. Yang, S. Zhu and Y. Lin: Regionlets for Generic Object Detection. T-PAMI 2015.
W. Zou, X. Wang, M. Sun and Y. Lin: Generic Object Detection with Dense Neural Patterns and Regionlets. British Machine Vision Conference 2014.
C. Long, X. Wang, G. Hua, M. Yang and Y. Lin: Accurate Object Detection with Location Relaxation and Regionlets Relocalization. Asian Conference on Computer Vision 2014.
42 vf-ssd 60.04 % 76.90 % 52.15 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
43 anm 59.21 % 75.51 % 56.49 % 3 s 1 core @ 2.5 Ghz (C/C++)
44 CompACT-Deep 58.73 % 69.70 % 52.69 % 1 s 1 core @ 2.5 Ghz (Matlab + C/C++)
Z. Cai, M. Saberian and N. Vasconcelos: Learning Complexity-Aware Cascades for Deep Pedestrian Detection. ICCV 2015.
45 DeepParts 58.68 % 70.46 % 52.73 % ~1 s GPU @ 2.5 Ghz (Matlab)
Y. Tian, P. Luo, X. Wang and X. Tang: Deep Learning Strong Parts for Pedestrian Detection. ICCV 2015.
46 AVOD-FPN
This method makes use of Velodyne laser scans.
code 58.42 % 67.32 % 57.44 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
47 LPN 58.18 % 70.54 % 54.18 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
48 RFCN 58.06 % 74.44 % 51.14 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
49 ACNet+BRFRes 57.23 % 67.60 % 51.79 % 0.55 s 1 core @ 2.5 Ghz (Matlab)
50 FilteredICF 57.12 % 69.05 % 51.46 % ~ 2 s >8 cores @ 2.5 Ghz (Matlab + C/C++)
S. Zhang, R. Benenson and B. Schiele: Filtered Channel Features for Pedestrian Detection. CVPR 2015.
51 p2dv 56.98 % 68.71 % 50.99 % 1 s 1 core @ 2.5 Ghz (C/C++)
52 FRCNN+Or code 56.78 % 71.18 % 52.86 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
53 D-TSF 56.77 % 69.03 % 50.77 % 1 s 1 core @ 2.5 Ghz (C/C++)
ERROR: Wrong syntax in BIBTEX file.
54 FD2 56.68 % 71.09 % 51.65 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
55 MV-RGBD-RF
This method makes use of Velodyne laser scans.
56.59 % 73.05 % 49.63 % 4 s 4 cores @ 2.5 Ghz (C/C++)
A. Gonzalez, D. Vazquez, A. Lopez and J. Amores: On-Board Object Detection: Multicue, Multimodal, and Multiview Random Forest of Local Experts.. IEEE Trans. on Cybernetics 2016.
A. Gonzalez, G. Villalonga, J. Xu, D. Vazquez, J. Amores and A. Lopez: Multiview Random Forest of Local Experts Combining RGB and LIDAR data for Pedestrian Detection. IEEE Intelligent Vehicles Symposium (IV) 2015.
56 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
55.74 % 65.73 % 49.08 % 0.06 s GPU @ 1.8 Ghz (Python)
57 YOLOv2-3cls 55.43 % 70.05 % 51.55 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
58 Vote3Deep
This method makes use of Velodyne laser scans.
55.38 % 67.94 % 52.62 % 1.5 s 4 cores @ 2.5 Ghz (C/C++)
M. Engelcke, D. Rao, D. Zeng Wang, C. Hay Tong and I. Posner: Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks. ArXiv e-prints 2016.
59 FD 55.33 % 67.87 % 50.02 % 0.01 s GPU @ >3.5 Ghz (Python)
60 pAUCEnsT 54.58 % 66.11 % 48.49 % 60 s 1 core @ 2.5 Ghz (Matlab + C/C++)
S. Paisitkriangkrai, C. Shen and A. Hengel: Pedestrian Detection with Spatially Pooled Features and Structured Ensemble Learning. arXiv 2014.
61 deprecated 54.02 % 70.43 % 49.83 % 0.00 s GPU @ 2.5 Ghz (C/C++)
62 PDV2 53.74 % 65.71 % 49.47 % 3.7 s 1 core @ 3.0 Ghz Matlab (C/C++)
J. Shen, X. Zuo, J. Li, W. Yang and H. Ling: A novel pixel neighborhood differential statistic feature for pedestrian and face detection . Pattern Recognition 2017.
63 CHTTL 53.04 % 66.26 % 49.48 % 0.07 s 1 core @ 2.5 Ghz (Python)
64 MTDP 52.97 % 66.97 % 47.64 % 0.15 s GPU @ 2.0 Ghz (Python)
65 FYSqueeze 52.60 % 66.07 % 48.40 % 0.01 s >8 cores @ 2.5 Ghz (Python)
66 HM 51.89 % 68.95 % 43.86 % 1 s 1 core @ 2.5 Ghz (C/C++)
67 Inha_cvlab 51.13 % 63.59 % 46.77 % 0.01 s GPU @ 2.5 Ghz (Python)
68 ACFD
This method makes use of Velodyne laser scans.
code 50.91 % 61.59 % 45.51 % 0.2 s 4 cores @ >3.5 Ghz (C/C++)
M. Dimitrievski, P. Veelaert and W. Philips: Semantically aware multilateral filter for depth upsampling in automotive LiDAR point clouds. IEEE Intelligent Vehicles Symposium, IV 2017, Los Angeles, CA, USA, June 11-14, 2017 2017.
69 R-CNN 50.20 % 62.05 % 44.85 % 4 s GPU @ 3.3 Ghz (C/C++)
J. Hosang, M. Omran, R. Benenson and B. Schiele: Taking a Deeper Look at Pedestrians. arXiv 2015.
70 SSD1 50.14 % 63.93 % 47.46 % 0.255 s GPU @ 2.5 Ghz (python+ C/C++)
71 tiny-det 47.81 % 62.02 % 45.53 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
72 fd3 47.67 % 59.28 % 44.46 % 0.01 s GPU @ 2.5 Ghz (C/C++)
73 ACF 47.29 % 60.11 % 42.90 % 1 s 1 core @ 3.5 Ghz (Matlab + C/C++)
P. Doll\'ar, R. Appel, S. Belongie and P. Perona: Fast Feature Pyramids for Object Detection. PAMI 2014.
74 Fusion-DPM
This method makes use of Velodyne laser scans.
code 46.67 % 59.38 % 42.05 % ~ 30 s 1 core @ 3.5 Ghz (Matlab + C/C++)
C. Premebida, J. Carreira, J. Batista and U. Nunes: Pedestrian Detection Combining RGB and Dense LIDAR Data. IROS 2014.
75 FSSPD 46.39 % 60.66 % 43.44 % 0.07 s GPU @ 2.0 Ghz (Python + C/C++)
76 ACF-MR 46.23 % 58.85 % 42.10 % 0.6 s 1 core @ 3.5 Ghz (C/C++)
R. Rajaram, E. Ohn-Bar and M. Trivedi: Looking at Pedestrians at Different Scales: A Multi-resolution Approach and Evaluations. T-ITS 2016.
77 HA-SSVM 45.51 % 58.91 % 41.08 % 21 s 1 core @ >3.5 Ghz (Matlab + C/C++)
J. Xu, S. Ramos, D. Vázquez and A. López: Hierarchical Adaptive Structural SVM for Domain Adaptation. IJCV 2016.
78 FRO 45.43 % 57.56 % 40.50 % 0.19 s GPU @ 2.5 Ghz (Python)
79 DPM-VOC+VP 44.86 % 59.60 % 40.37 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
80 ACF-SC 44.77 % 54.20 % 39.57 % <0.3 s 1 core @ >3.5 Ghz (Matlab + C/C++)
C. Cadena, A. Dick and I. Reid: A Fast, Modular Scene Understanding System using Context-Aware Object Detection. Robotics and Automation (ICRA), 2015 IEEE International Conference on 2015.
81 SquaresICF code 44.42 % 57.47 % 40.08 % 1 s GPU @ >3.5 Ghz (C/C++)
R. Benenson, M. Mathias, T. Tuytelaars and L. Gool: Seeking the strongest rigid detector. CVPR 2013.
82 LXT-DET 44.18 % 61.27 % 43.54 % 0.1 s 1 core @ 2.5 Ghz (Python)
83 VxNet(LiDAR)
This method makes use of Velodyne laser scans.
44.08 % 50.61 % 42.84 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
84 AR-FCN 43.88 % 53.16 % 35.58 % 0.19 s GPU @ 2.5 Ghz (C/C++)
85 AVOD
This method makes use of Velodyne laser scans.
code 43.49 % 51.64 % 37.79 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
86 YOLOv2 code 43.33 % 53.02 % 35.41 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
87 Roadstar.ai 42.72 % 47.02 % 42.36 % 0.08 s GPU @ 2.0 Ghz (Python)
88 GNN 42.56 % 58.22 % 40.53 % 0.2 s 1 core @ 2.5 Ghz (Python)
89 SubCat 42.34 % 54.06 % 37.95 % 1.2 s 6 cores @ 2.5 Ghz (Matlab + C/C++)
E. Ohn-Bar and M. Trivedi: Fast and Robust Object Detection Using Visual Subcategories. Computer Vision and Pattern Recognition Workshops Mobile Vision 2014.
90 bin 40.91 % 55.95 % 39.05 % 15ms s GPU @ >3.5 Ghz (Python)
91 GNN 40.69 % 55.22 % 38.65 % 0.2 s 1 core @ 2.5 Ghz (Python)
92 ACF 40.62 % 49.08 % 36.66 % 0.2 s 1 core @ >3.5 Ghz (Matlab + C/C++)
P. Doll\'ar, R. Appel, S. Belongie and P. Perona: Fast Feature Pyramids for Object Detection. PAMI 2014.
P. Doll\'ar: Piotr's Image and Video Matlab Toolbox (PMT). .
93 NMRDO 40.59 % 55.43 % 39.75 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
94 LSVM-MDPM-sv 39.36 % 51.75 % 35.95 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
95 LSVM-MDPM-us code 38.35 % 50.01 % 34.78 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
96 NLK 36.48 % 42.71 % 34.93 % 0.05 s 1 core @ 2.5 Ghz (Python + C/C++)
97 Vote3D
This method makes use of Velodyne laser scans.
35.74 % 44.47 % 33.72 % 0.5 s 4 cores @ 2.8 Ghz (C/C++)
D. Wang and I. Posner: Voting for Voting in Online Point Cloud Object Detection. Proceedings of Robotics: Science and Systems 2015.
98 3dSSD 34.86 % 44.59 % 34.77 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
99 BNet 34.40 % 41.05 % 28.88 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
100 mBoW
This method makes use of Velodyne laser scans.
31.37 % 44.36 % 30.62 % 10 s 1 core @ 2.5 Ghz (C/C++)
J. Behley, V. Steinhage and A. Cremers: Laser-based Segment Classification Using a Mixture of Bag-of-Words. Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2013.
101 BirdNet
This method makes use of Velodyne laser scans.
30.90 % 36.83 % 29.93 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
102 DPM-C8B1
This method uses stereo information.
29.03 % 38.96 % 25.61 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
103 YOLO 24.35 % 25.63 % 17.50 % 0.03 s GPU @ 1.0 Ghz (C/C++)
104 R-CNN_VGG 23.16 % 28.95 % 22.17 % 10 s GPU @ 2.5 Ghz (Matlab + C/C++)
105 LMNetV2
This method makes use of Velodyne laser scans.
23.00 % 27.03 % 23.26 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
106 Kyolo3 20.99 % 25.73 % 20.51 % 0.16 s 4 cores @ 2.5 Ghz (Python)
107 Fast-SSD 16.30 % 23.14 % 16.06 % 0.06 s GTX650Ti
108 YOLOv2 code 16.19 % 20.80 % 15.43 % 0.02 s GPU @ 3.5 Ghz (C/C++)
J. Redmon, S. Divvala, R. Girshick and A. Farhadi: You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
109 BIP-HETERO 13.38 % 14.85 % 13.25 % ~2 s 1 core @ 2.5 Ghz (C/C++)
A. Mekonnen, F. Lerasle, A. Herbulot and C. Briand: People Detection with Heterogeneous Features and Explicit Optimization on Computation Time. Pattern Recognition (ICPR), 2014 22nd International Conference on 2014.
110 TopNet-DecayRate
This method makes use of Velodyne laser scans.
10.84 % 13.06 % 11.03 % 0.12 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
111 TopNet-HighRes
This method makes use of Velodyne laser scans.
10.62 % 12.69 % 10.92 % 0.27 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
112 LMnetV1.1 6.96 % 7.96 % 6.94 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
113 LMnet
This method makes use of Velodyne laser scans.
6.50 % 7.34 % 6.57 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
Table as LaTeX | Only published Methods

Cyclist


Method Setting Code Moderate Easy Hard Runtime Environment
1 iDST-VT 78.21 % 86.06 % 69.47 % 1 s GPU @ 2.5 Ghz (C/C++)
2 SWC 77.58 % 86.02 % 68.44 % 0.5 s GPU @ >3.5 Ghz (Python + C/C++)
3 VCTNet 77.20 % 83.41 % 68.78 % 0.18 s GPU @ 3.5 GHz (C/C++)
4 RRC code 76.47 % 84.96 % 65.46 % 3.6 s GPU @ 2.5 Ghz (Python + C/C++)
J. Ren, X. Chen, J. Liu, W. Sun, J. Pang, Q. Yan, Y. Tai and L. Xu: Accurate Single Stage Detector Using Recurrent Rolling Convolution. CVPR 2017.
5 MS-CNN code 74.45 % 82.34 % 64.91 % 0.4 s GPU @ 2.5 Ghz (C/C++)
Z. Cai, Q. Fan, R. Feris and N. Vasconcelos: A Unified Multi-scale Deep Convolutional Neural Network for Fast Object Detection. ECCV 2016.
6 TuSimple code 74.26 % 81.38 % 64.88 % 1.6 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit all the layers: Fast and accurate cnn object detector with scale dependent pooling and cascaded rejection classifiers. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
K. He, X. Zhang, S. Ren and J. Sun: Deep residual learning for image recognition. Proceedings of the IEEE conference on computer vision and pattern recognition 2016.
7 Deep3DBox 73.48 % 82.65 % 64.11 % 1.5 s GPU @ 2.5 Ghz (C/C++)
A. Mousavian, D. Anguelov, J. Flynn and J. Kosecka: 3D Bounding Box Estimation Using Deep Learning and Geometry. CVPR 2017.
8 SiRtAKI 73.34 % 84.59 % 64.72 % 0.18 s GPU @ >3.5 Ghz (C/C++)
9 SDP+RPN 73.08 % 81.05 % 64.88 % 0.4 s GPU @ 2.5 Ghz (Python + C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in Neural Information Processing Systems 2015.
10 sensekitti code 72.50 % 81.76 % 64.00 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
11 F-PointNet
This method makes use of Velodyne laser scans.
code 72.25 % 84.90 % 65.14 % 0.17 s GPU @ 3.0 Ghz (Python)
C. Qi, W. Liu, C. Wu, H. Su and L. Guibas: Frustum PointNets for 3D Object Detection from RGB-D Data. arXiv preprint arXiv:1711.08488 2017.
12 RCL-FC 72.01 % 79.77 % 63.31 % 0.19 s GPU @ 2.5 Ghz (Matlab + C/C++)
13 SubCNN 70.77 % 77.82 % 62.71 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
14 Sogo_MM 70.72 % 77.57 % 62.23 % 1.5 s GPU @ 2.5 Ghz (C/C++)
15 3DOP
This method uses stereo information.
code 68.81 % 80.17 % 61.36 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
16 eCHIEV 68.38 % 82.17 % 60.59 % 1.5 s 1 core @ 2.5 Ghz (C/C++)
17 Pose-RCNN 68.04 % 80.19 % 59.95 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
18 Vote3Deep
This method makes use of Velodyne laser scans.
67.96 % 76.49 % 62.88 % 1.5 s 4 cores @ 2.5 Ghz (C/C++)
M. Engelcke, D. Rao, D. Zeng Wang, C. Hay Tong and I. Posner: Vote3Deep: Fast Object Detection in 3D Point Clouds Using Efficient Convolutional Neural Networks. ArXiv e-prints 2016.
19 IVA code 67.36 % 77.63 % 59.62 % 0.4 s GPU @ 2.5 Ghz (C/C++)
Y. Zhu, J. Wang, C. Zhao, H. Guo and H. Lu: Scale-adaptive Deconvolutional Regression Network for Pedestrian Detection. ACCV 2016.
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in neural information processing systems 2015.
20 DeepStereoOP 65.72 % 77.00 % 57.74 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
21 HSR2 64.94 % 76.36 % 57.62 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
22 Roadstar.ai 64.48 % 75.74 % 57.79 % 0.08 s GPU @ 2.0 Ghz (Python)
23 R-DML 63.90 % 76.60 % 56.98 % 0.16 s GPU @ 2.5 Ghz (Python + C/C++)
24 HM3D 63.89 % 76.28 % 56.51 % 0.35 s GPU @ >3.5 Ghz (C/C++)
25 Mono3D code 63.85 % 75.22 % 58.96 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
26 WRInception 62.85 % 78.19 % 55.64 % 0.06 s GPU @ 2.5 Ghz (C/C++)
27 Faster R-CNN code 62.81 % 71.41 % 55.44 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
S. Ren, K. He, R. Girshick and J. Sun: Faster R-CNN: Towards Real- Time Object Detection with Region Proposal Networks. NIPS 2015.
28 SDP+CRC (ft) 60.87 % 74.31 % 53.95 % 0.6 s GPU @ 2.5 Ghz (C/C++)
F. Yang, W. Choi and Y. Lin: Exploit All the Layers: Fast and Accurate CNN Object Detector with Scale Dependent Pooling and Cascaded Rejection Classifiers. Proceedings of the IEEE International Conference on Computer Vision and Pattern Recognition 2016.
29 VxNet(LiDAR)
This method makes use of Velodyne laser scans.
59.33 % 72.04 % 54.72 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
30 AVOD-FPN
This method makes use of Velodyne laser scans.
code 59.32 % 68.65 % 55.82 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
31 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
58.94 % 81.96 % 57.20 % 0.06 s GPU @ 1.8 Ghz (Python)
32 Regionlets 58.69 % 70.09 % 51.81 % 1 s >8 cores @ 2.5 Ghz (C/C++)
X. Wang, M. Yang, S. Zhu and Y. Lin: Regionlets for Generic Object Detection. T-PAMI 2015.
W. Zou, X. Wang, M. Sun and Y. Lin: Generic Object Detection with Dense Neural Patterns and Regionlets. British Machine Vision Conference 2014.
C. Long, X. Wang, G. Hua, M. Yang and Y. Lin: Accurate Object Detection with Location Relaxation and Regionlets Relocalization. Asian Conference on Computer Vision 2014.
33 FRCNN+Or code 57.37 % 70.05 % 51.00 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
34 VAT-Net 57.01 % 68.71 % 50.14 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
35 AVOD
This method makes use of Velodyne laser scans.
code 56.01 % 65.72 % 48.89 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
36 ReSqueeze 54.93 % 68.34 % 49.19 % 0.03 s GPU @ >3.5 Ghz (Python)
37 HNet code 54.10 % 69.71 % 48.02 % 1 s 1 core @ 2.5 Ghz (Python + C/C++)
38 deprecated 52.95 % 69.91 % 46.80 % 0.00 s GPU @ 2.5 Ghz (C/C++)
39 NLK 52.74 % 60.66 % 48.49 % 0.05 s 1 core @ 2.5 Ghz (Python + C/C++)
40 YOLOv2-3cls 51.67 % 68.14 % 45.79 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
41 RFCN 51.19 % 63.26 % 44.54 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
42 anm 50.54 % 67.40 % 45.22 % 3 s 1 core @ 2.5 Ghz (C/C++)
43 tiny-det 50.48 % 63.78 % 44.23 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
44 LPN 50.02 % 65.33 % 44.85 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
45 BirdNet
This method makes use of Velodyne laser scans.
49.04 % 64.88 % 46.61 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
46 FYSqueeze 48.80 % 67.03 % 43.82 % 0.01 s >8 cores @ 2.5 Ghz (Python)
47 vf-ssd 48.49 % 65.74 % 46.94 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
48 CHTTL 48.28 % 64.06 % 43.07 % 0.07 s 1 core @ 2.5 Ghz (Python)
49 FD2 44.29 % 62.32 % 40.65 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
50 MTDP 43.08 % 54.53 % 38.79 % 0.15 s GPU @ 2.0 Ghz (Python)
51 HM 42.99 % 60.32 % 41.45 % 1 s 1 core @ 2.5 Ghz (C/C++)
52 FRO 42.98 % 59.96 % 38.97 % 0.19 s GPU @ 2.5 Ghz (Python)
53 GNN 42.65 % 59.43 % 37.72 % 0.2 s 1 core @ 2.5 Ghz (Python)
54 MV-RGBD-RF
This method makes use of Velodyne laser scans.
42.61 % 51.46 % 37.42 % 4 s 4 cores @ 2.5 Ghz (C/C++)
A. Gonzalez, D. Vazquez, A. Lopez and J. Amores: On-Board Object Detection: Multicue, Multimodal, and Multiview Random Forest of Local Experts.. IEEE Trans. on Cybernetics 2016.
A. Gonzalez, G. Villalonga, J. Xu, D. Vazquez, J. Amores and A. Lopez: Multiview Random Forest of Local Experts Combining RGB and LIDAR data for Pedestrian Detection. IEEE Intelligent Vehicles Symposium (IV) 2015.
55 Inha_cvlab 42.39 % 60.04 % 38.26 % 0.01 s GPU @ 2.5 Ghz (Python)
56 AR-FCN 41.83 % 51.05 % 33.99 % 0.19 s GPU @ 2.5 Ghz (C/C++)
57 YOLOv2 code 39.96 % 56.59 % 33.06 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
58 BNet 38.07 % 54.91 % 30.98 % 0.03 s 1 core @ 2.5 Ghz (C/C++)
59 pAUCEnsT 37.88 % 52.28 % 33.38 % 60 s 1 core @ 2.5 Ghz (Matlab + C/C++)
S. Paisitkriangkrai, C. Shen and A. Hengel: Pedestrian Detection with Spatially Pooled Features and Structured Ensemble Learning. arXiv 2014.
60 GNN 37.64 % 54.47 % 35.09 % 0.2 s 1 core @ 2.5 Ghz (Python)
61 FD 37.01 % 51.41 % 32.93 % 0.01 s GPU @ >3.5 Ghz (Python)
62 3dSSD 34.00 % 42.51 % 32.04 % 0.03 s GPU @ 2.5 Ghz (Python + C/C++)
63 NMRDO 33.43 % 46.39 % 27.79 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
64 bin 31.60 % 44.33 % 28.12 % 15ms s GPU @ >3.5 Ghz (Python)
65 Vote3D
This method makes use of Velodyne laser scans.
31.24 % 41.45 % 28.60 % 0.5 s 4 cores @ 2.8 Ghz (C/C++)
D. Wang and I. Posner: Voting for Voting in Online Point Cloud Object Detection. Proceedings of Robotics: Science and Systems 2015.
66 DPM-VOC+VP 31.16 % 43.65 % 28.29 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
67 LSVM-MDPM-us code 30.81 % 40.31 % 28.17 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
68 LSVM-MDPM-sv 29.24 % 37.71 % 27.52 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
69 DPM-C8B1
This method uses stereo information.
29.04 % 43.28 % 26.20 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
70 R-CNN_VGG 28.79 % 37.71 % 25.82 % 10 s GPU @ 2.5 Ghz (Matlab + C/C++)
71 TopNet-HighRes
This method makes use of Velodyne laser scans.
22.83 % 29.88 % 20.11 % 0.27 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
72 mBoW
This method makes use of Velodyne laser scans.
21.62 % 28.19 % 20.93 % 10 s 1 core @ 2.5 Ghz (C/C++)
J. Behley, V. Steinhage and A. Cremers: Laser-based Segment Classification Using a Mixture of Bag-of-Words. Proc. of the IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) 2013.
73 TopNet-DecayRate
This method makes use of Velodyne laser scans.
18.28 % 27.57 % 18.62 % 0.12 s NVIDIA GeForce 1080 Ti (tensorflow-gpu)
S. Wirges, T. Fischer, J. Frias and C. Stiller: Object Detection and Classification in Occupancy Grid Maps using Deep Convolutional Networks. 2018.
74 YOLO 13.96 % 18.07 % 13.83 % 0.03 s GPU @ 1.0 Ghz (C/C++)
75 LMNetV2
This method makes use of Velodyne laser scans.
12.80 % 18.67 % 12.43 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
76 Kyolo3 9.09 % 9.09 % 9.09 % 0.16 s 4 cores @ 2.5 Ghz (Python)
77 Fast-SSD 7.10 % 11.77 % 7.23 % 0.06 s GTX650Ti
78 YOLOv2 code 4.55 % 4.55 % 4.55 % 0.02 s GPU @ 3.5 Ghz (C/C++)
J. Redmon, S. Divvala, R. Girshick and A. Farhadi: You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2016.
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
79 LMnetV1.1 2.61 % 1.91 % 2.73 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
80 LMnet
This method makes use of Velodyne laser scans.
1.73 % 2.13 % 1.99 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
Table as LaTeX | Only published Methods

Object Detection and Orientation Estimation Evaluation

Cars


Method Setting Code Moderate Easy Hard Runtime Environment
1 SAITv1 89.93 % 90.60 % 79.78 % 0.18 s GPU @ 2.5 Ghz (Python, C/C++)
2 Deep MANTA 89.86 % 97.19 % 80.39 % 0.7 s GPU @ 2.5 Ghz (Python + C/C++)
F. Chabot, M. Chaouch, J. Rabarisoa, C. Teulière and T. Chateau: Deep MANTA: A Coarse-to-fine Many-Task Network for joint 2D and 3D vehicle analysis from monocular image. CVPR 2017.
3 RaC 89.25 % 89.98 % 80.07 % 1s s GPU @ 1.0 Ghz (C/C++)
4 M3D 89.23 % 90.41 % 79.60 % 0.4 s GPU @ 2.5 Ghz (Python + C/C++)
5 Sogo_MM 88.72 % 90.67 % 78.95 % 1.5 s GPU @ 2.5 Ghz (C/C++)
6 Deep3DBox 88.56 % 90.39 % 77.17 % 1.5 s GPU @ 2.5 Ghz (C/C++)
A. Mousavian, D. Anguelov, J. Flynn and J. Kosecka: 3D Bounding Box Estimation Using Deep Learning and Geometry. CVPR 2017.
7 SubCNN 88.43 % 90.61 % 78.63 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
8 LTT 87.90 % 89.83 % 80.72 % 0.4 s 1 core @ 3.5 Ghz (Python)
9 AVOD
This method makes use of Velodyne laser scans.
code 87.46 % 89.59 % 79.54 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
10 HM3D 87.29 % 89.41 % 77.08 % 0.35 s GPU @ >3.5 Ghz (C/C++)
11 AVOD-FPN
This method makes use of Velodyne laser scans.
code 87.13 % 89.95 % 79.74 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
12 MonoFusion 87.03 % 90.35 % 76.37 % 0.12 s TITAN X GPU
13 DeepStereoOP 86.57 % 89.01 % 77.13 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
14 Mono3D code 85.83 % 89.00 % 76.00 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
15 3DOP
This method uses stereo information.
code 85.81 % 88.56 % 76.21 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
16 AVOD-SSD
This method makes use of Velodyne laser scans.
code 85.05 % 88.69 % 77.35 % 0.09 s GPU @ 2.5 Ghz (Python)
17 MBR-SSD 85.03 % 88.10 % 75.92 % 4.0 s GPU @ 2.5 Ghz (Python + C/C++)
18 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
81.31 % 87.84 % 71.95 % 0.06 s GPU @ 1.8 Ghz (Python)
19 FRCNN+Or code 77.61 % 88.52 % 67.69 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
20 avodC 76.30 % 86.31 % 68.71 % 0.1 s GPU @ 2.5 Ghz (Python)
21 MB-Net 76.12 % 85.38 % 59.84 % 0.02 s GPU @ 1.5 Ghz (C/C++)
22 3D FCN
This method makes use of Velodyne laser scans.
75.71 % 85.46 % 68.19 % >5 s 1 core @ 2.5 Ghz (C/C++)
B. Li: 3D Fully Convolutional Network for Vehicle Detection in Point Cloud. IROS 2017.
23 3D-SSMFCNN code 75.42 % 75.44 % 67.27 % 0.1 s GPU @ 1.5 Ghz (C/C++)
L. Novak: Vehicle Detection and Pose Estimation for Autonomous Driving. 2017.
24 Pose-RCNN 75.35 % 88.78 % 61.47 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
25 VS3D 75.16 % 83.52 % 59.59 % 0.58 s GPU @ 2.5 Ghz (C/C++)
26 3DVP code 74.59 % 81.02 % 64.11 % 40 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Data-Driven 3D Voxel Patterns for Object Category Recognition. IEEE Conference on Computer Vision and Pattern Recognition 2015.
27 SubCat code 74.42 % 80.74 % 58.83 % 0.7 s 6 cores @ 3.5 Ghz (Matlab + C/C++)
E. Ohn-Bar and M. Trivedi: Learning to Detect Vehicles by Clustering Appearance Patterns. T-ITS 2015.
28 BdCost48LDCF code 66.01 % 77.10 % 50.35 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
29 BdCost48-25C 65.25 % 77.59 % 50.68 % 4 s 1 core @ 2.5 Ghz (C/C++)
30 OC-DPM 64.88 % 74.66 % 52.24 % 10 s 8 cores @ 2.5 Ghz (Matlab)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Occlusion Patterns for Object Class Detection. IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 2013.
31 3DVSSD 64.72 % 77.22 % 57.56 % 0.06 s 1 core @ 2.5 Ghz (C/C++)
32 DPM-VOC+VP 63.27 % 77.51 % 47.57 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
33 AOG-View 62.25 % 77.37 % 50.44 % 3 s 1 core @ 2.5 Ghz (Matlab, C/C++)
B. Li, T. Wu and S. Zhu: Integrating Context and Occlusion for Car Detection by Hierarchical And-Or Model. ECCV 2014.
34 NMRDO 59.55 % 77.38 % 51.91 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
35 SAMME48LDCF code 57.49 % 75.12 % 46.64 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
36 LSVM-MDPM-sv 56.69 % 70.86 % 45.91 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
37 VeloFCN
This method makes use of Velodyne laser scans.
52.70 % 70.21 % 46.11 % 1 s GPU @ 2.5 Ghz (Python + C/C++)
B. Li, T. Zhang and T. Xia: Vehicle Detection from 3D Lidar Using Fully Convolutional Network. RSS 2016 .
38 VAT-Net 51.79 % 53.92 % 46.44 % 1 s GPU @ 2.5 Ghz (Python + C/C++)
39 desNet 51.78 % 54.13 % 46.46 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
40 DesNet 51.15 % 53.16 % 45.77 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
41 SN-net 51.12 % 52.55 % 45.89 % 0.02 s GPU @ 2.5 Ghz (Python + C/C++)
42 vfnet 50.96 % 53.53 % 45.91 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
43 DPM-C8B1
This method uses stereo information.
50.32 % 59.53 % 39.22 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
44 InNet 49.63 % 51.71 % 44.89 % 0.16 s GPU @ 3.5 Ghz (Python + C/C++)
45 denet 49.43 % 51.85 % 44.99 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
46 CNN-ds code 48.12 % 46.39 % 38.40 % 0.05 s 1 core @ 2.5 Ghz (Python)
47 HSR2 45.46 % 47.03 % 40.60 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
48 ReSqueeze 45.40 % 47.38 % 41.68 % 0.03 s GPU @ >3.5 Ghz (Python)
49 WRInception 45.07 % 47.05 % 40.52 % 0.06 s GPU @ 2.5 Ghz (C/C++)
50 VCTNet 44.78 % 48.36 % 40.04 % 0.18 s GPU @ 3.5 GHz (C/C++)
51 sensekitti code 44.56 % 47.06 % 41.50 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
52 vf-ssd 44.33 % 44.39 % 41.36 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
53 LMNetV2
This method makes use of Velodyne laser scans.
43.40 % 58.86 % 37.15 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
54 fd3 41.85 % 47.02 % 37.77 % 0.01 s GPU @ 2.5 Ghz (C/C++)
55 Inha_cvlab 41.63 % 46.83 % 37.38 % 0.01 s GPU @ 2.5 Ghz (Python)
56 FD 40.40 % 46.30 % 34.01 % 0.01 s GPU @ >3.5 Ghz (Python)
57 FD2 39.44 % 47.56 % 35.20 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
58 LMnetV1.1 36.66 % 52.94 % 30.28 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
59 LMnet
This method makes use of Velodyne laser scans.
35.86 % 50.81 % 29.90 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
60 BirdNet
This method makes use of Velodyne laser scans.
35.81 % 50.85 % 34.90 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
61 SDN
This method makes use of Velodyne laser scans.
34.48 % 47.13 % 31.40 % 0.096 s GPU @ 1.7 Ghz (Python)
62 YOLOv2-3cls 34.10 % 35.61 % 30.21 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
63 MMOD+CNN code 33.90 % 36.43 % 28.49 % 0.28 s 4 cores @ >3.5 Ghz (C/C++)
64 FYSqueeze 33.84 % 33.19 % 30.65 % 0.01 s >8 cores @ 2.5 Ghz (Python)
65 RFCN 33.59 % 37.75 % 29.82 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
66 deprecated 32.72 % 34.26 % 28.06 % 0.00 s GPU @ 2.5 Ghz (C/C++)
67 LPN 32.41 % 33.97 % 29.15 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
68 SceneNet 32.02 % 36.62 % 28.46 % 0.03 s GPU @ 2.5 Ghz (C/C++)
69 MTDP 31.04 % 34.12 % 27.50 % 0.15 s GPU @ 2.0 Ghz (Python)
70 AOG code 30.81 % 34.05 % 24.86 % 3 s 4 cores @ 2.5 Ghz (Matlab)
T. Wu, B. Li and S. Zhu: Learning And-Or Models to Represent Context and Occlusion for Car Detection and Viewpoint Estimation. TPAMI 2016.
B. Li, T. Wu and S. Zhu: Integrating Context and Occlusion for Car Detection by Hierarchical And-Or Model. ECCV 2014.
71 Fast-SSD 29.60 % 40.48 % 25.85 % 0.06 s GTX650Ti
72 Roadstar.ai 29.44 % 32.87 % 26.77 % 0.08 s GPU @ 2.0 Ghz (Python)
73 CHTTL 29.30 % 32.38 % 26.44 % 0.07 s 1 core @ 2.5 Ghz (Python)
74 YOLOv2 code 26.98 % 34.61 % 23.42 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
75 SubCat48LDCF code 26.78 % 34.43 % 19.46 % 0.5 s 8 cores @ 3.5 Ghz (Matlab + C/C++)
A. Fernández-Baldera, J. Buenaposada and L. Baumela: BAdaCost: Multi-class Boosting with Costs . Pattern Recognition 2018.
76 bin 25.79 % 31.13 % 22.90 % 15ms s GPU @ >3.5 Ghz (Python)
77 CSoR
This method makes use of Velodyne laser scans.
code 25.38 % 34.43 % 21.95 % 3.5 s 4 cores @ >3.5 Ghz (Python + C/C++)
L. Plotkin: PyDriver: Entwicklung eines Frameworks für räumliche Detektion und Klassifikation von Objekten in Fahrzeugumgebung. 2015.
78 RT3D
This method makes use of Velodyne laser scans.
18.98 % 24.23 % 20.56 % 0.09 s GPU @ 1.8Ghz
79 Kyolo3 18.21 % 19.50 % 15.99 % 0.16 s 4 cores @ 2.5 Ghz (Python)
80 VoxelNet basic
This method makes use of Velodyne laser scans.
18.12 % 18.64 % 16.65 % 0.07 s GPU (Python)
81 Licar
This method makes use of Velodyne laser scans.
15.58 % 18.24 % 16.15 % 0.09 s GPU @ 2.0 Ghz (Python)
82 DoBEM 14.02 % 15.35 % 16.33 % 0.6 s GPU @ 2.5 Ghz (Python + C/C++)
S. Yu, T. Westfechtel, R. Hamada, K. Ohno and S. Tadokoro: Vehicle Detection and Localization on Bird's Eye View Elevation Images Using Convolutional Neural Network. IEEE International Symposium on Safety, Security, and Rescue Robotics (SSRR) 2017.
83 SPC
This method makes use of Velodyne laser scans.
12.12 % 15.61 % 11.23 % 0.4 s 4 cores @ 2.5 Ghz (Python)
84 LidarNet
This method makes use of Velodyne laser scans.
1.09 % 0.70 % 0.88 % 0.007 s GPU @ 2.5 Ghz (C/C++)
Table as LaTeX | Only published Methods


Pedestrians


Method Setting Code Moderate Easy Hard Runtime Environment
1 Sogo_MM 66.83 % 78.89 % 62.06 % 1.5 s GPU @ 2.5 Ghz (C/C++)
2 SubCNN 66.28 % 78.33 % 61.37 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
3 Pose-RCNN 59.89 % 74.10 % 54.21 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
4 3DOP
This method uses stereo information.
code 59.79 % 73.46 % 57.04 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
5 DeepStereoOP 59.28 % 73.37 % 56.87 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
6 HM3D 58.21 % 70.22 % 53.72 % 0.35 s GPU @ >3.5 Ghz (C/C++)
7 Mono3D code 58.12 % 68.58 % 54.94 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
8 FRCNN+Or code 52.62 % 66.84 % 48.72 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
9 AVOD-FPN
This method makes use of Velodyne laser scans.
code 44.92 % 53.36 % 43.77 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
10 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
43.51 % 51.56 % 38.78 % 0.06 s GPU @ 1.8 Ghz (Python)
11 DPM-VOC+VP 39.83 % 53.66 % 35.73 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
12 VCTNet 38.73 % 42.30 % 36.69 % 0.18 s GPU @ 3.5 GHz (C/C++)
13 sensekitti code 37.50 % 43.55 % 35.08 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
14 AVOD
This method makes use of Velodyne laser scans.
code 36.38 % 44.12 % 31.81 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
15 LSVM-MDPM-sv 35.49 % 47.00 % 32.42 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
16 RFCN 35.26 % 44.20 % 31.60 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
17 WRInception 35.14 % 40.34 % 32.50 % 0.06 s GPU @ 2.5 Ghz (C/C++)
18 SubCat 34.18 % 43.95 % 30.76 % 1.2 s 6 cores @ 2.5 Ghz (Matlab + C/C++)
E. Ohn-Bar and M. Trivedi: Fast and Robust Object Detection Using Visual Subcategories. Computer Vision and Pattern Recognition Workshops Mobile Vision 2014.
19 YOLOv2-3cls 33.97 % 42.38 % 31.83 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
20 HSR2 33.86 % 39.97 % 32.48 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
21 NMRDO 33.06 % 44.95 % 31.83 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
22 SSD1 32.73 % 41.73 % 30.69 % 0.255 s GPU @ 2.5 Ghz (python+ C/C++)
23 FYSqueeze 32.66 % 40.20 % 30.25 % 0.01 s >8 cores @ 2.5 Ghz (Python)
24 RPN+BF code 32.55 % 40.97 % 29.52 % 0.6 s GPU @ 2.5 Ghz (Matlab + C/C++)
L. Zhang, L. Lin, X. Liang and K. He: Is Faster R-CNN Doing Well for Pedestrian Detection?. ECCV 2016.
25 ReSqueeze 32.35 % 37.95 % 30.38 % 0.03 s GPU @ >3.5 Ghz (Python)
26 LPN 31.63 % 38.40 % 28.90 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
27 VAT-Net 31.45 % 35.14 % 28.10 % 0.08 s GPU @ 2.5 Ghz (Python)
28 deprecated 30.04 % 39.60 % 27.56 % 0.00 s GPU @ 2.5 Ghz (C/C++)
29 MTDP 29.04 % 36.90 % 25.96 % 0.15 s GPU @ 2.0 Ghz (Python)
30 CHTTL 29.01 % 36.41 % 26.95 % 0.07 s 1 core @ 2.5 Ghz (Python)
31 FD2 28.59 % 35.53 % 26.02 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
32 ACF 28.46 % 35.69 % 26.18 % 1 s 1 core @ 3.5 Ghz (Matlab + C/C++)
P. Doll\'ar, R. Appel, S. Belongie and P. Perona: Fast Feature Pyramids for Object Detection. PAMI 2014.
33 vf-ssd 28.37 % 35.28 % 24.76 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
34 FD 27.90 % 33.68 % 25.17 % 0.01 s GPU @ >3.5 Ghz (Python)
35 YOLOv2 code 27.35 % 32.98 % 22.99 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
36 Inha_cvlab 26.96 % 33.08 % 24.74 % 0.01 s GPU @ 2.5 Ghz (Python)
37 bin 26.22 % 34.76 % 25.12 % 15ms s GPU @ >3.5 Ghz (Python)
38 fd3 25.38 % 31.30 % 24.18 % 0.01 s GPU @ 2.5 Ghz (C/C++)
39 DPM-C8B1
This method uses stereo information.
23.37 % 31.08 % 20.72 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
40 LXT-DET 23.22 % 30.35 % 23.00 % 0.1 s 1 core @ 2.5 Ghz (Python)
41 ACF-MR 23.18 % 29.35 % 21.00 % 0.6 s 1 core @ 3.5 Ghz (C/C++)
R. Rajaram, E. Ohn-Bar and M. Trivedi: Looking at Pedestrians at Different Scales: A Multi-resolution Approach and Evaluations. T-ITS 2016.
42 Roadstar.ai 22.18 % 24.51 % 21.91 % 0.08 s GPU @ 2.0 Ghz (Python)
43 BirdNet
This method makes use of Velodyne laser scans.
17.26 % 21.34 % 16.67 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
44 LMNetV2
This method makes use of Velodyne laser scans.
14.79 % 18.18 % 14.67 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
45 Kyolo3 9.67 % 12.06 % 9.27 % 0.16 s 4 cores @ 2.5 Ghz (Python)
46 Fast-SSD 9.16 % 12.68 % 9.01 % 0.06 s GTX650Ti
47 LMnetV1.1 4.51 % 5.28 % 4.39 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
48 LMnet
This method makes use of Velodyne laser scans.
4.13 % 4.71 % 4.09 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
Table as LaTeX | Only published Methods


Cyclists


Method Setting Code Moderate Easy Hard Runtime Environment
1 Sogo_MM 63.59 % 70.70 % 56.15 % 1.5 s GPU @ 2.5 Ghz (C/C++)
2 SubCNN 63.41 % 71.39 % 56.34 % 2 s GPU @ 3.5 Ghz (Python + C/C++)
Y. Xiang, W. Choi, Y. Lin and S. Savarese: Subcategory-aware Convolutional Neural Networks for Object Proposals and Detection. IEEE Winter Conference on Applications of Computer Vision (WACV) 2017.
3 Pose-RCNN 62.25 % 74.85 % 55.09 % 2 s >8 cores @ 2.5 Ghz (Python)
M. Braun, Q. Rao, Y. Wang and F. Flohr: Pose-RCNN: Joint object detection and pose estimation using 3D object proposals. Intelligent Transportation Systems (ITSC), 2016 IEEE 19th International Conference on 2016.
4 Deep3DBox 59.37 % 68.58 % 51.97 % 1.5 s GPU @ 2.5 Ghz (C/C++)
A. Mousavian, D. Anguelov, J. Flynn and J. Kosecka: 3D Bounding Box Estimation Using Deep Learning and Geometry. CVPR 2017.
5 3DOP
This method uses stereo information.
code 58.59 % 71.95 % 52.35 % 3s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Y. Zhu, A. Berneshawi, H. Ma, S. Fidler and R. Urtasun: 3D Object Proposals for Accurate Object Class Detection. NIPS 2015.
6 AVOD-FPN
This method makes use of Velodyne laser scans.
code 57.53 % 67.61 % 54.16 % 0.1 s Titan X (Pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
7 VxNet++(LiDAR)
This method makes use of Velodyne laser scans.
57.20 % 80.97 % 55.14 % 0.06 s GPU @ 1.8 Ghz (Python)
8 DeepStereoOP 55.62 % 67.49 % 48.85 % 3.4 s GPU @ 3.5 Ghz (Matlab + C/C++)
C. Pham and J. Jeon: Robust Object Proposals Re-ranking for Object Detection in Autonomous Driving Using Convolutional Neural Networks. Signal Processing: Image Communiation 2017.
9 HM3D 55.12 % 67.32 % 48.86 % 0.35 s GPU @ >3.5 Ghz (C/C++)
10 AVOD
This method makes use of Velodyne laser scans.
code 54.43 % 64.36 % 47.67 % 0.08 s Titan X (pascal)
J. Ku, M. Mozifian, J. Lee, A. Harakeh and S. Waslander: Joint 3D Proposal Generation and Object Detection from View Aggregation. arXiv preprint arXiv:1712.02294 2017.
11 Mono3D code 53.11 % 65.74 % 48.87 % 4.2 s GPU @ 2.5 Ghz (Matlab + C/C++)
X. Chen, K. Kundu, Z. Zhang, H. Ma, S. Fidler and R. Urtasun: Monocular 3D Object Detection for Autonomous Driving. CVPR 2016.
12 FRCNN+Or code 50.91 % 63.41 % 45.46 % 0.09 s Titan Xp GPU
C. Guindel, D. Martin and J. Armingol: Joint Object Detection and Viewpoint Estimation using CNN features. IEEE International Conference on Vehicular Electronics and Safety (ICVES) 2017.
13 VCTNet 43.14 % 48.22 % 38.67 % 0.18 s GPU @ 3.5 GHz (C/C++)
14 sensekitti code 42.12 % 46.65 % 36.66 % 4.5 s GPU @ 2.5 Ghz (Python + C/C++)
B. Yang, J. Yan, Z. Lei and S. Li: Craft Objects from Images. CVPR 2016.
15 HSR2 36.82 % 42.76 % 32.33 % 0.15 s 1 core @ 2.5 Ghz (C/C++)
16 WRInception 34.02 % 41.88 % 29.37 % 0.06 s GPU @ 2.5 Ghz (C/C++)
17 VAT-Net 32.94 % 40.36 % 28.04 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
18 BirdNet
This method makes use of Velodyne laser scans.
30.76 % 41.48 % 28.66 % 0.11 s Titan Xp GPU
J. Beltran, C. Guindel, F. Moreno, D. Cruzado, F. Garcia and A. Escalera: BirdNet: a 3D Object Detection Framework from LiDAR information. arXiv preprint arXiv:1805.01195 2018.
19 vf-ssd 27.67 % 38.92 % 26.24 % 0.08 s GPU @ 2.5 Ghz (Python + C/C++)
20 Roadstar.ai 27.58 % 32.23 % 24.88 % 0.08 s GPU @ 2.0 Ghz (Python)
21 ReSqueeze 27.40 % 35.39 % 24.32 % 0.03 s GPU @ >3.5 Ghz (Python)
22 LPN 27.01 % 32.96 % 25.01 % 0.2 s GPU @ 2.5 Ghz (Python + C/C++)
23 RFCN 26.92 % 32.01 % 23.86 % 0.3 s GPU @ 1.0 Ghz (Python + C/C++)
24 FYSqueeze 26.71 % 34.53 % 24.39 % 0.01 s >8 cores @ 2.5 Ghz (Python)
25 FD2 24.65 % 35.58 % 21.97 % 0.01 s GPU @ >3.5 Ghz (Python + C/C++)
26 deprecated 24.05 % 31.01 % 21.12 % 0.00 s GPU @ 2.5 Ghz (C/C++)
27 NMRDO 23.53 % 32.68 % 19.81 % 0.1 s GPU @ 2.5 Ghz (Python + C/C++)
28 CHTTL 23.51 % 30.24 % 21.05 % 0.07 s 1 core @ 2.5 Ghz (Python)
29 Inha_cvlab 23.38 % 33.62 % 20.73 % 0.01 s GPU @ 2.5 Ghz (Python)
30 DPM-VOC+VP 23.22 % 31.24 % 21.62 % 8 s 1 core @ 2.5 Ghz (C/C++)
B. Pepik, M. Stark, P. Gehler and B. Schiele: Multi-view and 3D Deformable Part Models. IEEE Transactions on Pattern Analysis and Machine Intelligence (TPAMI) 2015.
31 LSVM-MDPM-sv 23.14 % 28.89 % 22.28 % 10 s 4 cores @ 3.0 Ghz (C/C++)
P. Felzenszwalb, R. Girshick, D. McAllester and D. Ramanan: Object Detection with Discriminatively Trained Part-Based Models. PAMI 2010.
A. Geiger, C. Wojek and R. Urtasun: Joint 3D Estimation of Objects and Scene Layout. NIPS 2011.
32 YOLOv2 code 22.36 % 28.97 % 19.45 % 0.03 s GPU @ 2.0 Ghz (Python + C/C++)
33 YOLOv2-3cls 22.03 % 28.27 % 19.80 % 0.05 s GPU @ 2.5 Ghz (C/C++)
J. Redmon and A. Farhadi: YOLO9000: Better, Faster, Stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2017.
34 FD 21.60 % 30.76 % 18.56 % 0.01 s GPU @ >3.5 Ghz (Python)
35 DPM-C8B1
This method uses stereo information.
19.25 % 27.16 % 17.95 % 15 s 4 cores @ 2.5 Ghz (Matlab + C/C++)
J. Yebes, L. Bergasa and M. García-Garrido: Visual Object Recognition with 3D-Aware Features in KITTI Urban Scenes. Sensors 2015.
J. Yebes, L. Bergasa, R. Arroyo and A. Lázaro: Supervised learning and evaluation of KITTI's cars detector with DPM. IV 2014.
36 MTDP 18.95 % 23.33 % 17.24 % 0.15 s GPU @ 2.0 Ghz (Python)
37 bin 12.64 % 17.86 % 11.33 % 15ms s GPU @ >3.5 Ghz (Python)
38 LMNetV2
This method makes use of Velodyne laser scans.
8.33 % 13.37 % 8.14 % 0.02 s GPU @ 2.5 Ghz (C/C++)
K. Minemura, H. Liau, A. Monrroy and S. Kato: LMNet: Real-time Multiclass Object Detection on CPU using 3D LiDAR. 2018.
39 Fast-SSD 4.55 % 6.94 % 4.55 % 0.06 s GTX650Ti
40 Kyolo3 3.96 % 3.96 % 3.96 % 0.16 s 4 cores @ 2.5 Ghz (Python)
41 LMnetV1.1 1.45 % 1.33 % 1.58 % 0.01 s GPU @ 1.0 Ghz (Python + C/C++)
42 LMnet
This method makes use of Velodyne laser scans.
1.16 % 1.61 % 1.39 % 0.013 s GPU @ 1.1 Ghz (Python + C/C++)
Table as LaTeX | Only published Methods


Related Datasets

Citation

When using this dataset in your research, we will be happy if you cite us:
@INPROCEEDINGS{Geiger2012CVPR,
  author = {Andreas Geiger and Philip Lenz and Raquel Urtasun},
  title = {Are we ready for Autonomous Driving? The KITTI Vision Benchmark Suite},
  booktitle = {Conference on Computer Vision and Pattern Recognition (CVPR)},
  year = {2012}
}



eXTReMe Tracker