Abstract
Most existing multi-source domain adaptation (MSDA) methods minimize the distance between multiple source-target domain pairs via feature distribution alignment, an approach borrowed from the single source setting. However, with diverse source domains, aligning pairwise feature distributions is challenging and could even be counterproductive for MSDA. In this paper, we introduce a novel approach: transferable attribute learning. The motivation is simple: although different domains can have drastically different visual appearances, they contain the same set of classes characterized by the same set of attributes; an MSDA model thus should focus on learning the most transferable attributes for the target domain. Adopting this approach, we propose a domain attention consistency network, dubbed DAC-Net. The key design is a feature channel attention module, which aims to identify transferable features (attributes). Importantly, the attention module is supervised by a consistency loss, which is imposed on the distributions of channel attention weights between source and target domains. Moreover, to facilitate discriminative feature learning on the target data, we combine pseudo-labeling with a class compactness loss to minimize the distance between the target features and the classifier's weight vectors. Extensive experiments on three MSDA benchmarks show that our DAC-Net achieves new state of the art performance on all of them.
Original language | English |
---|---|
Pages | 1-13 |
Number of pages | 13 |
Publication status | Published - Nov 2021 |
Event | 32nd British Machine Vision Conference, BMVC 2021 - Virtual, Online Duration: 22 Nov 2021 → 25 Nov 2021 https://www.bmvc2021-virtualconference.com/ https://www.bmvc2021-virtualconference.com/programme/accepted-papers/ |
Conference
Conference | 32nd British Machine Vision Conference, BMVC 2021 |
---|---|
Period | 22/11/21 → 25/11/21 |
Internet address |
Scopus Subject Areas
- Artificial Intelligence
- Computer Vision and Pattern Recognition