TY - GEN
T1 - FA3D: Fast and Accurate 3D Object Detection
AU - Habibi Aghdam, Hamed
AU - Demilew, Selameab
AU - Laganiére, Robert
AU - Petriu, Emil
PY - 2020/12/7
Y1 - 2020/12/7
N2 - Fast and accurate detection of objects, in 3D, is one of the critical components in an advanced driver assistance system. In this paper, we aim to develop an accurate 3D object detector that runs in near real-time on low-end embedded systems. We propose an efficient framework that converts raw point cloud into a 3D occupancy cuboid and detects cars using a deep convolutional neural network. Even though the complexity of our proposed model is high, it runs at 7.27 FPS on a Jetson Xavier and at 57.83 FPS on a high-end workstation that is 18 % and 43 % faster than the fastest published method while having a comparable performance with state-of-the-art models on the KITTI dataset. We conduct a comprehensive error analysis on our model and show that two quantities are the principal sources of error among nine predicted attributes. Our source code is available at https://github.com/Selameab/FA3D. © 2020, Springer Nature Switzerland AG.
AB - Fast and accurate detection of objects, in 3D, is one of the critical components in an advanced driver assistance system. In this paper, we aim to develop an accurate 3D object detector that runs in near real-time on low-end embedded systems. We propose an efficient framework that converts raw point cloud into a 3D occupancy cuboid and detects cars using a deep convolutional neural network. Even though the complexity of our proposed model is high, it runs at 7.27 FPS on a Jetson Xavier and at 57.83 FPS on a high-end workstation that is 18 % and 43 % faster than the fastest published method while having a comparable performance with state-of-the-art models on the KITTI dataset. We conduct a comprehensive error analysis on our model and show that two quantities are the principal sources of error among nine predicted attributes. Our source code is available at https://github.com/Selameab/FA3D. © 2020, Springer Nature Switzerland AG.
KW - 3D object detection
KW - Deep neural networks
KW - Smart and autonomous vehicles
UR - http://www.scopus.com/inward/record.url?scp=85098134042&partnerID=8YFLogxK
U2 - 10.1007/978-3-030-64556-4_31
DO - 10.1007/978-3-030-64556-4_31
M3 - Conference contribution
SN - 9783030645557
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 397
EP - 409
BT - Advances in Visual Computing - 15th International Symposium, ISVC 2020, Proceedings
A2 - Bebis, George
A2 - Yin, Zhaozheng
A2 - Kim, Edward
A2 - Bender, Jan
A2 - Subr, Kartic
A2 - Kwon, Bum Chul
A2 - Zhao, Jian
A2 - Kalkofen, Denis
A2 - Baciu, George
ER -