Image Re-Identification: Where Self-supervision Meets Vision-Language Learning
This work addresses image re-identification for computer vision applications, presenting an incremental improvement by combining existing self-supervision and vision-language learning techniques.
The paper tackles image re-identification by integrating self-supervision with pre-trained CLIP models, proposing SVLL-ReID, which achieves state-of-the-art performance on six benchmark datasets without using concrete text labels.
Recently, large-scale vision-language pre-trained models like CLIP have shown impressive performance in image re-identification (ReID). In this work, we explore whether self-supervision can aid in the use of CLIP for image ReID tasks. Specifically, we propose SVLL-ReID, the first attempt to integrate self-supervision and pre-trained CLIP via two training stages to facilitate the image ReID. We observe that: 1) incorporating language self-supervision in the first training stage can make the learnable text prompts more distinguishable, and 2) incorporating vision self-supervision in the second training stage can make the image features learned by the image encoder more discriminative. These observations imply that: 1) the text prompt learning in the first stage can benefit from the language self-supervision, and 2) the image feature learning in the second stage can benefit from the vision self-supervision. These benefits jointly facilitate the performance gain of the proposed SVLL-ReID. By conducting experiments on six image ReID benchmark datasets without any concrete text labels, we find that the proposed SVLL-ReID achieves the overall best performances compared with state-of-the-arts. Codes will be publicly available at https://github.com/BinWangGzhu/SVLL-ReID.