MILA: Memory-Based Instance-Level Adaptation for Cross-Domain Object Detection
Onkar Krishna (Hitachi Ltd.),* Hiroki Ohashi (Hitachi Ltd), Saptarshi Sinha (University of Bristol)
The 34th British Machine Vision Conference
Abstract
Cross-domain object detection is challenging, and it involves aligning labeled source and unlabeled target domains. Previous approaches have used adversarial training to align features at both image-level and instance-level. At the instance level, finding a suitable source sample that aligns with a target sample is crucial. A source sample is considered suitable if it differs from the target sample only in domain, without differences in unimportant characteristics such as orientation and color, which can hinder the model’s focus on aligning the domain difference. However, existing instance-level feature alignment methods struggle to find suitable source instances because their search scope is limited to mini-batches. The insufficient diversity of mini-batches becomes problematic particularly when the target instances have high intra-class variance. To address this issue, we propose a memory-based instance-level domain adaptation framework. Our method aligns a target instance with the most similar source instance of the same category retrieved from a memory storage. Specifically, we introduce a memory module that dynamically stores the pooled features of all labeled source instances, categorized by their labels. Additionally, we introduce a simple yet effective memory retrieval module that retrieves a set of matching memory slots for target instances. Our experiments on various domain shift scenarios demonstrate that our approach outperforms existing non-memory-based methods significantly. Code is available at https://github.com/hitachi-rd-cv/MILAVideo
Citation
@inproceedings{Krishna_2023_BMVC,
author = {Onkar Krishna and Hiroki Ohashi and Saptarshi Sinha},
title = {MILA: Memory-Based Instance-Level Adaptation for Cross-Domain Object Detection},
booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023},
publisher = {BMVA},
year = {2023},
url = {https://papers.bmvc2023.org/0192.pdf}
}
Copyright © 2023
The British Machine Vision Conference is organised by