10 April 2018 Border-oriented post-processing refinement on detected vehicle bounding box for ADAS
Author Affiliations +
Proceedings Volume 10615, Ninth International Conference on Graphic and Image Processing (ICGIP 2017); 106150B (2018) https://doi.org/10.1117/12.2305330
Event: Ninth International Conference on Graphic and Image Processing, 2017, Qingdao, China
Abstract
We investigate a new approach for improving localization accuracy of detected vehicles for object detection in advanced driver assistance systems(ADAS). Specifically, we implement a bounding box refinement as a post-processing of the state-of-the-art object detectors (Faster R-CNN, YOLOv2, etc.). The bounding box refinement is achieved by individually adjusting each border of the detected bounding box to its target location using a regression method. We use HOG features which perform well on the edge detection of vehicles to train the regressor and the regressor is independent of the CNN-based object detectors. Experiment results on the KITTI 2012 benchmark show that we can achieve up to 6% improvements over YOLOv2 and Faster R-CNN object detectors on the IoU threshold of 0.8. Also, the proposed refinement framework is computationally light, allowing for processing one bounding box within a few milliseconds on CPU. Further, this refinement method can be added to any object detectors, especially those with high speed but less accuracy.
© (2018) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Xinyuan Chen, Xinyuan Chen, Zhaoning Zhang, Zhaoning Zhang, Minne Li, Minne Li, Dongsheng Li, Dongsheng Li, "Border-oriented post-processing refinement on detected vehicle bounding box for ADAS", Proc. SPIE 10615, Ninth International Conference on Graphic and Image Processing (ICGIP 2017), 106150B (10 April 2018); doi: 10.1117/12.2305330; https://doi.org/10.1117/12.2305330
PROCEEDINGS
9 PAGES


SHARE
RELATED CONTENT

An improved corner detection algorithm for image sequence
Proceedings of SPIE (November 23 2014)
Part identification in robotic assembly using vision system
Proceedings of SPIE (December 23 2013)
Fast scene layout estimation via deep hashing
Proceedings of SPIE (July 25 2018)
Multirobot learning in an inherently cooperative task
Proceedings of SPIE (September 19 2001)
Use of mission roles as a robotic tasking device
Proceedings of SPIE (November 14 1999)

Back to Top