Submission details of TANet++

Name TANet++
Paper Link https://arxiv.org/pdf/2106.15366.pdf
Code Link https://github.com/happinesslz/TANet
AP 0.63922
Input N/A
Runtime 0.28 s
Environment 1 Titan Tesla K40c
Abstract TANet is one of state-of-the-art 3D object detection method on KITTI and JRDB benchmark, the network contains a Triple Attention module and Coarse-to-Fine Regression module to improve the robustness and accuracy of 3D Detection. However, since the original input data (point clouds) contains a lot of noise during collecting the data, which will further affect the training of the model. For example, the object is far from the robot, the sensor is difficult to obtain enough pointcloud. If the objects only contains few point clouds, and the samples are fed into model with the normal samples together during training, the detector will be difficult to distinguish the individual with few pointcloud belong to object or background. In this paper, we propose TANet++ to improve the performance on 3D Detection, which adopt a novel training strategy on training the TANet. In order to reduce the negative impact by the weak samples, the training strategy previously filtered the training data, and then the TANet++ is trained by the rest of data. The experimental results shows that AP score of TANet++ is 8.98% higher than TANet on JRDB benchmark.

Detailed results

Overall precision/recall curve

Per-sequence results

sequence name AP
cubberly-auditorium-2019-04-22_1 0.709646
discovery-walk-2019-02-28_0 0.658285
discovery-walk-2019-02-28_1 0.691434
food-trucks-2019-02-12_0 0.745069
gates-ai-lab-2019-04-17_0 0.661827
gates-basement-elevators-2019-01-17_0 0.805537
gates-foyer-2019-01-17_0 0.824157
gates-to-clark-2019-02-28_0 0.75401
hewlett-class-2019-01-23_0 0.819215
hewlett-class-2019-01-23_1 0.915634
huang-2-2019-01-25_1 0.65003
huang-intersection-2019-01-22_0 0.626508
indoor-coupa-cafe-2019-02-06_0 0.537171
lomita-serra-intersection-2019-01-30_0 0.700028
meyer-green-2019-03-16_1 0.518292
nvidia-aud-2019-01-25_0 0.50542
nvidia-aud-2019-04-18_1 0.576044
nvidia-aud-2019-04-18_2 0.674196
outdoor-coupa-cafe-2019-02-06_0 0.537497
quarry-road-2019-02-28_0 0.701916
serra-street-2019-01-30_0 0.623736
stlc-111-2019-04-19_1 0.893132
stlc-111-2019-04-19_2 0.834973
tressider-2019-03-16_2 0.698366
tressider-2019-04-26_0 0.614667
tressider-2019-04-26_1 0.669851
tressider-2019-04-26_3 0.614804
total 0.63922