Search papers, labs, and topics across Lattice.
This paper introduces Prompt-Anchored vision-text Distillation (PAD) for lifelong person re-identification (LReID), addressing semantic drift, limited adaptability, and catastrophic forgetting in sequentially trained models. PAD leverages a frozen text encoder from pretrained vision-language models as a stable semantic anchor across domains, distilling prompts to preserve vision-text alignment. Results demonstrate that PAD outperforms state-of-the-art methods in both seen and unseen domains, balancing stability and plasticity.
Freezing a text encoder and distilling prompts from vision-language models can stabilize semantics and boost performance in lifelong person re-identification, even across unseen domains.
Lifelong person re-identification (LReID) aims to train a generalizable model with sequentially collected data. However, such models often suffer from semantic drift, limited adaptability, and catastrophic forgetting as new domains emerge. Existing exemplar-free approaches largely rely on visual-only distillation or parameter regularization, while overlooking the potential of auxiliary modalities, such as text, to preserve semantic stability and enable incremental plasticity. We observe that the frozen text encoder in pretrained vision-language models can serve as a stable semantic anchor across domains. To decouple the roles of vision and text, we propose Prompt-Anchored vision-text Distillation (PAD), an asymmetric vision-text framework for semantic alignment and cross-domain generalization. On the textual side, we distill prompts to preserve vision-text alignment under a fixed semantic space, acting as a global semantic reference rather than a dominant learning signal. On the visual side, an EMA-based teacher with an adaptive prompt pool enables domain-wise adaptation by allocating new slots while freezing past ones. Extensive experiments show that PAD substantially outperforms state-of-the-art methods across seen and unseen domains, achieving a strong balance between stability and plasticity. Project page is available at https://github.com/zu-zi/PAD.
