Anti-Forgetting Adaptation for Unsupervised Person Re-identification
This addresses the issue of model forgetting in incremental domain adaptation for person re-identification, which is important for real-world applications where data domains change over time, but it is incremental as it builds on existing adaptation methods.
The paper tackles the problem of forgetting previously learned knowledge in unsupervised domain adaptive person re-identification when adapting to new domains, proposing a framework that uses prototype and instance-level consistency to mitigate forgetting and improve generalization, with experiments showing significant improvements in anti-forgetting and generalization abilities.
Regular unsupervised domain adaptive person re-identification (ReID) focuses on adapting a model from a source domain to a fixed target domain. However, an adapted ReID model can hardly retain previously-acquired knowledge and generalize to unseen data. In this paper, we propose a Dual-level Joint Adaptation and Anti-forgetting (DJAA) framework, which incrementally adapts a model to new domains without forgetting source domain and each adapted target domain. We explore the possibility of using prototype and instance-level consistency to mitigate the forgetting during the adaptation. Specifically, we store a small number of representative image samples and corresponding cluster prototypes in a memory buffer, which is updated at each adaptation step. With the buffered images and prototypes, we regularize the image-to-image similarity and image-to-prototype similarity to rehearse old knowledge. After the multi-step adaptation, the model is tested on all seen domains and several unseen domains to validate the generalization ability of our method. Extensive experiments demonstrate that our proposed method significantly improves the anti-forgetting, generalization and backward-compatible ability of an unsupervised person ReID model.