Domain Adaptive YOLO for One-Stage Cross-Domain Detection

Shizhao Zhang, Hongya Tuo, Jian Hu, Zhongliang Jing
Proceedings of The 13th Asian Conference on Machine Learning, PMLR 157:785-797, 2021.

Abstract

Domain shift is a major challenge for object detectors to generalize well to real world applications. Emerging techniques of domain adaptation for two-stage detectors help to tackle this problem. However, two-stage detectors are not the first choice for industrial applications due to its long time consumption. In this paper, a novel Domain Adaptive YOLO (DA-YOLO) is proposed to improve cross-domain performance for one-stage detectors. Image level features alignment is used to strictly match for local features like texture, and loosely match for global features like illumination. Multi-scale instance level features alignment is presented to reduce instance domain shift effectively, such as variations in object appearance and viewpoint. A consensus regularization to these domain classifiers is employed to help the network generate domain-invariant detections. We evaluate our proposed method on popular datasets like Cityscapes, KITTI, SIM10K and et al.. The results demonstrate considerable improvement when tested under different cross-domain scenarios.

Cite this Paper

BibTeX
@InProceedings{pmlr-v157-zhang21c, title = {Domain Adaptive YOLO for One-Stage Cross-Domain Detection}, author = {Zhang, Shizhao and Tuo, Hongya and Hu, Jian and Jing, Zhongliang}, booktitle = {Proceedings of The 13th Asian Conference on Machine Learning}, pages = {785--797}, year = {2021}, editor = {Balasubramanian, Vineeth N. and Tsang, Ivor}, volume = {157}, series = {Proceedings of Machine Learning Research}, month = {17--19 Nov}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v157/zhang21c/zhang21c.pdf}, url = {https://proceedings.mlr.press/v157/zhang21c.html}, abstract = {Domain shift is a major challenge for object detectors to generalize well to real world applications. Emerging techniques of domain adaptation for two-stage detectors help to tackle this problem. However, two-stage detectors are not the first choice for industrial applications due to its long time consumption. In this paper, a novel Domain Adaptive YOLO (DA-YOLO) is proposed to improve cross-domain performance for one-stage detectors. Image level features alignment is used to strictly match for local features like texture, and loosely match for global features like illumination. Multi-scale instance level features alignment is presented to reduce instance domain shift effectively, such as variations in object appearance and viewpoint. A consensus regularization to these domain classifiers is employed to help the network generate domain-invariant detections. We evaluate our proposed method on popular datasets like Cityscapes, KITTI, SIM10K and et al.. The results demonstrate considerable improvement when tested under different cross-domain scenarios.} }
Endnote
%0 Conference Paper %T Domain Adaptive YOLO for One-Stage Cross-Domain Detection %A Shizhao Zhang %A Hongya Tuo %A Jian Hu %A Zhongliang Jing %B Proceedings of The 13th Asian Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2021 %E Vineeth N. Balasubramanian %E Ivor Tsang %F pmlr-v157-zhang21c %I PMLR %P 785--797 %U https://proceedings.mlr.press/v157/zhang21c.html %V 157 %X Domain shift is a major challenge for object detectors to generalize well to real world applications. Emerging techniques of domain adaptation for two-stage detectors help to tackle this problem. However, two-stage detectors are not the first choice for industrial applications due to its long time consumption. In this paper, a novel Domain Adaptive YOLO (DA-YOLO) is proposed to improve cross-domain performance for one-stage detectors. Image level features alignment is used to strictly match for local features like texture, and loosely match for global features like illumination. Multi-scale instance level features alignment is presented to reduce instance domain shift effectively, such as variations in object appearance and viewpoint. A consensus regularization to these domain classifiers is employed to help the network generate domain-invariant detections. We evaluate our proposed method on popular datasets like Cityscapes, KITTI, SIM10K and et al.. The results demonstrate considerable improvement when tested under different cross-domain scenarios.
APA
Zhang, S., Tuo, H., Hu, J. & Jing, Z.. (2021). Domain Adaptive YOLO for One-Stage Cross-Domain Detection. Proceedings of The 13th Asian Conference on Machine Learning, in Proceedings of Machine Learning Research 157:785-797 Available from https://proceedings.mlr.press/v157/zhang21c.html.

Related Material