Well annotated dataset is crucial to the training of object detectors. However, the production of finely annotated datasets for object detection tasks is extremely labor-intensive, therefore, cloud sourcing is often used to create datasets, which leads to these datasets tending to contain incorrect annotations such as inaccurate localization bounding boxes. In this study, we highlight a problem of object detection with noisy bounding box annotations and show that these noisy annotations are harmful to the performance of deep neural networks. To solve this problem, we further propose a framework to allow the network to modify the noisy datasets by alternating refinement. The experimental results demonstrate that our proposed framework can significantly alleviate the influences of noise on model performance.
Jiafeng MAO
The University of Tokyo
Qing YU
The University of Tokyo
Kiyoharu AIZAWA
The University of Tokyo
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copy
Jiafeng MAO, Qing YU, Kiyoharu AIZAWA, "Noisy Localization Annotation Refinement for Object Detection" in IEICE TRANSACTIONS on Information,
vol. E104-D, no. 9, pp. 1478-1485, September 2021, doi: 10.1587/transinf.2021EDP7026.
Abstract: Well annotated dataset is crucial to the training of object detectors. However, the production of finely annotated datasets for object detection tasks is extremely labor-intensive, therefore, cloud sourcing is often used to create datasets, which leads to these datasets tending to contain incorrect annotations such as inaccurate localization bounding boxes. In this study, we highlight a problem of object detection with noisy bounding box annotations and show that these noisy annotations are harmful to the performance of deep neural networks. To solve this problem, we further propose a framework to allow the network to modify the noisy datasets by alternating refinement. The experimental results demonstrate that our proposed framework can significantly alleviate the influences of noise on model performance.
URL: https://global.ieice.org/en_transactions/information/10.1587/transinf.2021EDP7026/_p
Copy
@ARTICLE{e104-d_9_1478,
author={Jiafeng MAO, Qing YU, Kiyoharu AIZAWA, },
journal={IEICE TRANSACTIONS on Information},
title={Noisy Localization Annotation Refinement for Object Detection},
year={2021},
volume={E104-D},
number={9},
pages={1478-1485},
abstract={Well annotated dataset is crucial to the training of object detectors. However, the production of finely annotated datasets for object detection tasks is extremely labor-intensive, therefore, cloud sourcing is often used to create datasets, which leads to these datasets tending to contain incorrect annotations such as inaccurate localization bounding boxes. In this study, we highlight a problem of object detection with noisy bounding box annotations and show that these noisy annotations are harmful to the performance of deep neural networks. To solve this problem, we further propose a framework to allow the network to modify the noisy datasets by alternating refinement. The experimental results demonstrate that our proposed framework can significantly alleviate the influences of noise on model performance.},
keywords={},
doi={10.1587/transinf.2021EDP7026},
ISSN={1745-1361},
month={September},}
Copy
TY - JOUR
TI - Noisy Localization Annotation Refinement for Object Detection
T2 - IEICE TRANSACTIONS on Information
SP - 1478
EP - 1485
AU - Jiafeng MAO
AU - Qing YU
AU - Kiyoharu AIZAWA
PY - 2021
DO - 10.1587/transinf.2021EDP7026
JO - IEICE TRANSACTIONS on Information
SN - 1745-1361
VL - E104-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2021
AB - Well annotated dataset is crucial to the training of object detectors. However, the production of finely annotated datasets for object detection tasks is extremely labor-intensive, therefore, cloud sourcing is often used to create datasets, which leads to these datasets tending to contain incorrect annotations such as inaccurate localization bounding boxes. In this study, we highlight a problem of object detection with noisy bounding box annotations and show that these noisy annotations are harmful to the performance of deep neural networks. To solve this problem, we further propose a framework to allow the network to modify the noisy datasets by alternating refinement. The experimental results demonstrate that our proposed framework can significantly alleviate the influences of noise on model performance.
ER -