Detection Challenge Results

Detection challenge results are presented below. The plot results of detectors can be found here.

The Precision-Recall curve of the submission can be drawn compared to the state-of-the-art detectors here.

Method Overall Easy Medium Hard Cloudy Night Rainy Sunny Speed Environment
1 SSD_VDIG 82.68% 94.60% 89.71% 70.65% 89.81% 83.02% 73.35% 88.11% 2 fps GPU:Titan X * 2
Anonymous submission
2 HAVD 80.51% 94.48% 86.13% 69.02% 87.28% 82.30% 69.37% 89.71% 2.1 fps CPU:2x Intel Xeon E5-2650v4 2.4GHZ, GPU:Titan X
THU CV-AI Lab
3 DCN 79.85% 93.85% 85.07% 69.00% 85.55% 82.38% 68.95% 89.08% 8 fps (Python) CPU:2x Intel Xeon E5-2650v4 2.4GHZ, GPU:Titan X
THU CV-AI Lab
4 HAT 78.64% 93.44% 83.09% 68.04% 86.27% 78.00% 67.97% 88.78% 3.6 fps (C++,Python) GPU:Titan X
Anonymous submission
5 GP-FRCNNm 77.96% 92.74% 82.39% 67.22% 83.23% 77.75% 70.17% 86.56% 4 fps (C++, Python) CPU:Intel(R) Xeon(R) CPU E5-2690v3 2.6GHZ, RAM:256GB, GPU:K40
Sikandar Amin, Fabio Galasso. Geometric Proposals for Faster R-CNN. AVSS 2017.
6 RTN 74.15% 91.52% 79.16% 61.73% 77.02% 77.20% 65.27% 84.14% 19.61 fps (C++) 2x Intel Xeon E5-2620v4, RAM:128GB,GPU:GTX1080
Anonymous submission
7 HPNDFCN 71.56% 93.51% 78.00% 55.62% 78.84% 79.37% 57.22% 84.92% 5 fps Intel(R) i5 2 core 2.10GHz, RAM:128GB,GPU:GTX1080Ti
Anonymous submission
8 R-FCN1 69.87% 93.32% 75.67% 54.31% 74.38% 75.09% 56.21% 84.08% 6 fps (Python) 2x Intel Xeon E5-2650v4, RAM:128GB,GPU:TitanX
Dai, J., Li, Y., He, K., & Sun, J. R-fcn: Object detection via region-based fully convolutional networks. In Advances in neural information processing systems(NIPS), 2016
9 ASDN-EB 69.60% 89.87% 75.09% 55.69% 75.27% 74.62% 54.85% 84.79% 7.69 fps CPU: 2x Intel Xeon E5-2650v4 2.4GHZ, GPU: Titan X
Anonymous submission
10 EB 67.96% 89.65% 73.12% 53.64% 72.42% 73.93% 53.40% 83.73% 10.00 fps (C++) 1x GPU:TitanX
Li Wang, Yao Lu, Hong Wang, Yingbin Zheng, Hao Ye, Xiangyang Xue, Evolving Boxes for Fast Vehicle Detection. In IEEE International Conference on Multimedia and Expo (ICME), 2017.
11 LateralCNN 67.25% 89.56% 73.59% 51.61% 69.11% 74.36% 55.77% 78.66% 21.47 fps Intel(R) i7-6800K 6 core 3.40GHz, RAM:32GB,GPU:TitanX
Anonymous submission
12 NANO 63.01% 80.33% 68.04% 50.73% 67.00% 62.20% 55.89% 73.89% - -
Anonymous submission
13 FasterRCNN2 58.45% 82.75% 63.05% 44.25% 66.29% 69.85% 45.16% 62.34% 11.11 fps (C++) 1x GPU:TitanX
Anonymous submission
14 YOLO2 57.72% 83.28% 62.25% 42.44% 57.97% 64.53% 47.84% 69.75% - 2x Intel Xeon E5-2620v4, RAM:128GB,GPU:GTX1080
Anonymous submission
15 CompACT 53.23% 64.84% 58.70% 43.16% 63.23% 46.37% 44.21% 71.16% 0.22 fps (Matlab,C++) 2x Intel Xeon E5-2470v2 @2.40GHz, RAM:64GB, GPU:Tesla K40
Z. Cai, M. Saberian, and N. Vasconcelos. Learning complexity-aware cascades for deep pedestrian detection. In ICCV, 2015. [code]
16 R-CNN 48.95% 59.31% 54.06% 39.47% 59.73% 39.32% 39.06% 67.52% 0.10 fps (Matlab,C++) 2x Intel Xeon E5-2470v2 @2.40GHz, RAM:64GB, GPU:Tesla K40
R. B. Girshick, J. Donahue, T. Darrell, and J. Malik. Rich feature hierarchies for accurate object detection and semantic segmentation. In CVPR, pages 580-587, 2014. [code]
17 ACF 46.35% 54.27% 51.52% 38.07% 58.30% 35.29% 37.09% 66.58% 0.67 fps (Matlab,C++) 2x Intel Xeon E5-2470v2 @2.40GHz, RAM:64GB
P. Dollár, R. Appel, S. Belongie, and P. Perona. Fast feature pyramids for object detection. In TPAMI, 36(8):1532-1545, 2014. [code]
18 SA-FRCNN 45.83% 73.93% 49.00% 30.76% 49.97% 52.30% 33.39% 55.04% - -
Anonymous submission
19 DPM 25.70% 34.42% 30.29% 17.62% 24.78% 30.91% 25.55% 31.77% 0.17 fps (Matlab,C++) 4x Intel Core i7-6600U @2.60GHz, RAM:8GB
P. F. Felzenszwalb, R. B. Girshick, D. A. McAllester, and D. Ramanan. Object detection with discriminatively trained part-based models. In TPAMI, 32(9):1627-1645, 2010. [code]