FA3D: Fast and Accurate 3D Object Detection

Hamed Habibi Aghdam, Selameab Demilew*, Robert Laganiére, Emil Petriu

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference paperpeer-review

Abstract

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.
Original languageEnglish
Title of host publicationAdvances in Visual Computing - 15th International Symposium, ISVC 2020, Proceedings
EditorsGeorge Bebis, Zhaozheng Yin, Edward Kim, Jan Bender, Kartic Subr, Bum Chul Kwon, Jian Zhao, Denis Kalkofen, George Baciu
Pages397-409
Number of pages13
DOIs
Publication statusPublished - 7 Dec 2020
Externally publishedYes
Event15th International Symposium on Visual Computing: ISVC 2020 - Virtuell, United States
Duration: 5 Oct 20207 Oct 2020

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume12509 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th International Symposium on Visual Computing
Country/TerritoryUnited States
CityVirtuell
Period5/10/207/10/20

Keywords

  • 3D object detection
  • Deep neural networks
  • Smart and autonomous vehicles

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this