Please use this identifier to cite or link to this item:
https://hdl.handle.net/10316/111956
Title: | Probabilistic Approach for Road-Users Detection | Authors: | Melotti, Gledson Lu, Weihao Conde, Pedro Zhao, Dezong Asvadi, Alireza Gonçalves, Nuno Premebida, Cristiano |
Keywords: | Object Detection; Overconfident prediction; Probabilistic calibration; Multimodality; Deep learning | Issue Date: | 2023 | Publisher: | IEEE | Serial title, monograph or event: | IEEE Transactions on Intelligent Transportation Systems | Volume: | 24 | Issue: | 9 | Abstract: | Object detection in autonomous driving applications implies the detection and tracking of semantic objects that are commonly native to urban driving environments, as pedestrians and vehicles. One of the major challenges in state-of-the-art deep-learning based object detection are false positives which occur with overconfident scores. This is highly undesirable in autonomous driving and other critical robotic-perception domains because of safety concerns. This paper proposes an approach to alleviate the problem of overconfident predictions by introducing a novel probabilistic layer to deep object detection networks in testing. The suggested approach avoids the traditional Sigmoid or Softmax prediction layer which often produces overconfident predictions. It is demonstrated that the proposed technique reduces overconfidence in the false positives without degrading the performance on the true positives. The approach is validated on the 2D-KITTI objection detection through the YOLOV4 and SECOND (Lidar-based detector). The proposed approach enables interpretable probabilistic predictions without the requirement of re-training the network and therefore is very practical. | URI: | https://hdl.handle.net/10316/111956 | ISSN: | 1524-9050 1558-0016 |
DOI: | 10.1109/TITS.2023.3268578 | Rights: | openAccess |
Appears in Collections: | FCTUC Eng.Electrotécnica - Artigos em Revistas Internacionais I&D ISR - Artigos em Revistas Internacionais |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Probabilistic Approach for Road-Users Detection.pdf | 6.87 MB | Adobe PDF | View/Open |
Page view(s)
56
checked on Aug 14, 2024
Download(s)
33
checked on Aug 14, 2024
Google ScholarTM
Check
Altmetric
Altmetric
This item is licensed under a Creative Commons License