Cargando…

Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving

Object detection plays a critical role in autonomous driving, but current state-of-the-art object detectors will inevitably fail in many driving scenes, which is unacceptable for safety-critical automated vehicles. Given the complexity of the real traffic scenarios, it is impractical to guarantee ze...

Descripción completa

Detalles Bibliográficos
Autores principales: Yang, Qinghua, Chen, Hui, Chen, Zhe, Su, Junzhe
Formato: Online Artículo Texto
Lenguaje:English
Publicado: MDPI 2021
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8073889/
https://www.ncbi.nlm.nih.gov/pubmed/33923776
http://dx.doi.org/10.3390/s21082819
_version_ 1783684233528606720
author Yang, Qinghua
Chen, Hui
Chen, Zhe
Su, Junzhe
author_facet Yang, Qinghua
Chen, Hui
Chen, Zhe
Su, Junzhe
author_sort Yang, Qinghua
collection PubMed
description Object detection plays a critical role in autonomous driving, but current state-of-the-art object detectors will inevitably fail in many driving scenes, which is unacceptable for safety-critical automated vehicles. Given the complexity of the real traffic scenarios, it is impractical to guarantee zero detection failure; thus, online failure prediction is of crucial importance to mitigate the risk of traffic accidents. Of all the failure cases, False Negative (FN) objects are most likely to cause catastrophic consequences, but little attention has been paid to the online FN prediction. In this paper, we propose a general introspection framework that can make online prediction of FN objects for black-box object detectors. In contrast to existing methods which rely on empirical assumptions or handcrafted features, we facilitate the FN feature extraction by an introspective FN predictor we designed in this framework. For this purpose, we extend the original concept of introspection to object-wise FN predictions, and propose a multi-branch cooperation mechanism to address the distinct foreground-background imbalance problem of FN objects. The effectiveness of the proposed framework is verified through extensive experiments and analysis, and the results show that our method successfully predicts the FN objects with 81.95% precision for 88.10% recall on the challenging KITTI Benchmark, and effectively improves object detection performance by taking FN predictions into consideration.
format Online
Article
Text
id pubmed-8073889
institution National Center for Biotechnology Information
language English
publishDate 2021
publisher MDPI
record_format MEDLINE/PubMed
spelling pubmed-80738892021-04-27 Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving Yang, Qinghua Chen, Hui Chen, Zhe Su, Junzhe Sensors (Basel) Article Object detection plays a critical role in autonomous driving, but current state-of-the-art object detectors will inevitably fail in many driving scenes, which is unacceptable for safety-critical automated vehicles. Given the complexity of the real traffic scenarios, it is impractical to guarantee zero detection failure; thus, online failure prediction is of crucial importance to mitigate the risk of traffic accidents. Of all the failure cases, False Negative (FN) objects are most likely to cause catastrophic consequences, but little attention has been paid to the online FN prediction. In this paper, we propose a general introspection framework that can make online prediction of FN objects for black-box object detectors. In contrast to existing methods which rely on empirical assumptions or handcrafted features, we facilitate the FN feature extraction by an introspective FN predictor we designed in this framework. For this purpose, we extend the original concept of introspection to object-wise FN predictions, and propose a multi-branch cooperation mechanism to address the distinct foreground-background imbalance problem of FN objects. The effectiveness of the proposed framework is verified through extensive experiments and analysis, and the results show that our method successfully predicts the FN objects with 81.95% precision for 88.10% recall on the challenging KITTI Benchmark, and effectively improves object detection performance by taking FN predictions into consideration. MDPI 2021-04-16 /pmc/articles/PMC8073889/ /pubmed/33923776 http://dx.doi.org/10.3390/s21082819 Text en © 2021 by the authors. https://creativecommons.org/licenses/by/4.0/Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
spellingShingle Article
Yang, Qinghua
Chen, Hui
Chen, Zhe
Su, Junzhe
Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title_full Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title_fullStr Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title_full_unstemmed Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title_short Introspective False Negative Prediction for Black-Box Object Detectors in Autonomous Driving
title_sort introspective false negative prediction for black-box object detectors in autonomous driving
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8073889/
https://www.ncbi.nlm.nih.gov/pubmed/33923776
http://dx.doi.org/10.3390/s21082819
work_keys_str_mv AT yangqinghua introspectivefalsenegativepredictionforblackboxobjectdetectorsinautonomousdriving
AT chenhui introspectivefalsenegativepredictionforblackboxobjectdetectorsinautonomousdriving
AT chenzhe introspectivefalsenegativepredictionforblackboxobjectdetectorsinautonomousdriving
AT sujunzhe introspectivefalsenegativepredictionforblackboxobjectdetectorsinautonomousdriving