Parameter-Efficient Vision-Language Adaptation with Continuous Metadata Conditioning for Animal Re-Identification

2026-07-10

AI Summary

This work addresses the challenge of identity–time distribution shift in long-term animal re-identification caused by morphological evolution and seasonal variations. To this end, the authors propose a CLIP-based, parameter-efficient vision–language adaptation framework featuring a novel continuous metadata conditioning mechanism that directly incorporates numerical attributes—such as age and season—into prompt representations, preserving their inherent continuity to smoothly modulate the embedding space and avoid information loss from discretization. The approach integrates low-rank visual adaptation, prompt supervision, and cross-modal alignment, leveraging metadata during training while requiring only visual inputs at inference. Evaluated on a seven-year longitudinal fish dataset and multiple wildlife benchmarks, the method consistently outperforms existing approaches under closed-set, open-set, and temporally aware protocols, demonstrating enhanced robustness to long-term appearance changes and temporal distribution shifts.

Abstract

Long-term animal re-identification (ReID) must remain robust to gradual morphological evolution and seasonal appearance shifts. Although recent vision-language models provide strong pretrained visual representations, adapting them to longitudinal ecological settings remains challenging, particularly under identity and temporal distribution shifts. We present a parameter-efficient CLIP adaptation framework for animal ReID and introduce a continuous metadata-conditioning mechanism that incorporates numerical attributes directly into the prompt representation during training. While low-rank visual adaptation, prompt-based supervision, and cross-modal alignment provide the adaptation framework, the proposed metadata-conditioning strategy constitutes the primary methodological contribution. By preserving the continuous structure of numerical metadata rather than discretizing it into textual categories, the proposed approach enables smooth modulation of the embedding space during training while maintaining a purely visual inference pipeline. Experiments on a seven-year longitudinal fish dataset and multiple wildlife benchmarks demonstrate improved performance under closed-set, open-set, and time-aware evaluation protocols. The results demonstrate that continuous metadata conditioning improves robustness to longitudinal appearance variation and temporal distribution shifts, while parameter-efficient adaptation enables a purely visual inference pipeline without requiring metadata at test time. Code and evaluation splits can be found at: https://github.com/AnilOsmanTur/MetaPrompt-ReID.

Problem

• animal re-identification
• longitudinal appearance variation
• temporal distribution shift
• vision-language models
• metadata conditioning

Innovation

• continuous metadata conditioning
• parameter-efficient adaptation
• vision-language models
• animal re-identification
• prompt-based learning