Distribution-Aware Calibration for Object Detection with Noisy Bounding Boxes

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Zhou, Donghao, Li, Jialin, Li, Jinpeng, Huang, Jiancheng, Nie, Qiang, Liu, Yong, Gao, Bin-Bin, Wang, Qiong, Heng, Pheng-Ann, Chen, Guangyong
Natura: Preprint
Pubblicazione: 2023
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910577054449664
author Zhou, Donghao
Li, Jialin
Li, Jinpeng
Huang, Jiancheng
Nie, Qiang
Liu, Yong
Gao, Bin-Bin
Wang, Qiong
Heng, Pheng-Ann
Chen, Guangyong
author_facet Zhou, Donghao
Li, Jialin
Li, Jinpeng
Huang, Jiancheng
Nie, Qiang
Liu, Yong
Gao, Bin-Bin
Wang, Qiong
Heng, Pheng-Ann
Chen, Guangyong
contents Large-scale well-annotated datasets are of great importance for training an effective object detector. However, obtaining accurate bounding box annotations is laborious and demanding. Unfortunately, the resultant noisy bounding boxes could cause corrupt supervision signals and thus diminish detection performance. Motivated by the observation that the real ground-truth is usually situated in the aggregation region of the proposals assigned to a noisy ground-truth, we propose DIStribution-aware CalibratiOn (DISCO) to model the spatial distribution of proposals for calibrating supervision signals. In DISCO, spatial distribution modeling is performed to statistically extract the potential locations of objects. Based on the modeled distribution, three distribution-aware techniques, i.e., distribution-aware proposal augmentation (DA-Aug), distribution-aware box refinement (DA-Ref), and distribution-aware confidence estimation (DA-Est), are developed to improve classification, localization, and interpretability, respectively. Extensive experiments on large-scale noisy image datasets (i.e., Pascal VOC and MS-COCO) demonstrate that DISCO can achieve state-of-the-art detection performance, especially at high noise levels. Code is available at https://github.com/Correr-Zhou/DISCO.
format Preprint
id arxiv_https___arxiv_org_abs_2308_12017
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Distribution-Aware Calibration for Object Detection with Noisy Bounding Boxes
Zhou, Donghao
Li, Jialin
Li, Jinpeng
Huang, Jiancheng
Nie, Qiang
Liu, Yong
Gao, Bin-Bin
Wang, Qiong
Heng, Pheng-Ann
Chen, Guangyong
Computer Vision and Pattern Recognition
Large-scale well-annotated datasets are of great importance for training an effective object detector. However, obtaining accurate bounding box annotations is laborious and demanding. Unfortunately, the resultant noisy bounding boxes could cause corrupt supervision signals and thus diminish detection performance. Motivated by the observation that the real ground-truth is usually situated in the aggregation region of the proposals assigned to a noisy ground-truth, we propose DIStribution-aware CalibratiOn (DISCO) to model the spatial distribution of proposals for calibrating supervision signals. In DISCO, spatial distribution modeling is performed to statistically extract the potential locations of objects. Based on the modeled distribution, three distribution-aware techniques, i.e., distribution-aware proposal augmentation (DA-Aug), distribution-aware box refinement (DA-Ref), and distribution-aware confidence estimation (DA-Est), are developed to improve classification, localization, and interpretability, respectively. Extensive experiments on large-scale noisy image datasets (i.e., Pascal VOC and MS-COCO) demonstrate that DISCO can achieve state-of-the-art detection performance, especially at high noise levels. Code is available at https://github.com/Correr-Zhou/DISCO.
title Distribution-Aware Calibration for Object Detection with Noisy Bounding Boxes
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2308.12017