Submission details of Team_HJ

Name Team_HJ
Paper Link N/A
Code Link N/A
AP 0.673813
Input N/A
Runtime 0.07 s
Environment 1 GPU (Titan X)
Abstract Panoramic image detection and tracking using both image and point clouds. For detection, we adopt the Cascade rcnn as our object detector. The backbone of the detector is ResNet 101. The detector is pre-trained on the COCO dataset and then fine tuned on the JRDB dataset. During training, we employ data augment methods such as mixup, random crop, multiscale training to augment the dataset. During inference, we employ horizontal flip and softnms techniques to generate more robust resutls.

Detailed results

Overall precision/recall curve

Per-sequence results

sequence name AP
cubberly-auditorium-2019-04-22_1 0.750113
discovery-walk-2019-02-28_0 0.825894
discovery-walk-2019-02-28_1 0.78908
food-trucks-2019-02-12_0 0.770223
gates-ai-lab-2019-04-17_0 0.743764
gates-basement-elevators-2019-01-17_0 0.899166
gates-foyer-2019-01-17_0 0.880852
gates-to-clark-2019-02-28_0 0.794834
hewlett-class-2019-01-23_0 0.899794
hewlett-class-2019-01-23_1 0.94646
huang-2-2019-01-25_1 0.635411
huang-intersection-2019-01-22_0 0.731425
indoor-coupa-cafe-2019-02-06_0 0.658213
lomita-serra-intersection-2019-01-30_0 0.678567
meyer-green-2019-03-16_1 0.501898
nvidia-aud-2019-01-25_0 0.638988
nvidia-aud-2019-04-18_1 0.741961
nvidia-aud-2019-04-18_2 0.824491
outdoor-coupa-cafe-2019-02-06_0 0.564904
quarry-road-2019-02-28_0 0.672068
serra-street-2019-01-30_0 0.621595
stlc-111-2019-04-19_1 0.844966
stlc-111-2019-04-19_2 0.860709
tressider-2019-03-16_2 0.635097
tressider-2019-04-26_0 0.598482
tressider-2019-04-26_1 0.66692
tressider-2019-04-26_3 0.647298
total 0.673813