Training-free Conditional Image Embedding Framework Leveraging Large Vision Language Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915694668414976 |
|---|---|
| author | Kawarada, Masayuki Yamada, Kosuke Tejero-de-Pablos, Antonio Inoue, Naoto |
| author_facet | Kawarada, Masayuki Yamada, Kosuke Tejero-de-Pablos, Antonio Inoue, Naoto |
| contents | Conditional image embeddings are feature representations that focus on specific aspects of an image indicated by a given textual condition (e.g., color, genre), which has been a challenging problem. Although recent vision foundation models, such as CLIP, offer rich representations of images, they are not designed to focus on a specified condition. In this paper, we propose DIOR, a method that leverages a large vision-language model (LVLM) to generate conditional image embeddings. DIOR is a training-free approach that prompts the LVLM to describe an image with a single word related to a given condition. The hidden state vector of the LVLM's last token is then extracted as the conditional image embedding. DIOR provides a versatile solution that can be applied to any image and condition without additional training or task-specific priors. Comprehensive experimental results on conditional image similarity tasks demonstrate that DIOR outperforms existing training-free baselines, including CLIP. Furthermore, DIOR achieves superior performance compared to methods that require additional training across multiple settings. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_21860 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Training-free Conditional Image Embedding Framework Leveraging Large Vision Language Models Kawarada, Masayuki Yamada, Kosuke Tejero-de-Pablos, Antonio Inoue, Naoto Computer Vision and Pattern Recognition Conditional image embeddings are feature representations that focus on specific aspects of an image indicated by a given textual condition (e.g., color, genre), which has been a challenging problem. Although recent vision foundation models, such as CLIP, offer rich representations of images, they are not designed to focus on a specified condition. In this paper, we propose DIOR, a method that leverages a large vision-language model (LVLM) to generate conditional image embeddings. DIOR is a training-free approach that prompts the LVLM to describe an image with a single word related to a given condition. The hidden state vector of the LVLM's last token is then extracted as the conditional image embedding. DIOR provides a versatile solution that can be applied to any image and condition without additional training or task-specific priors. Comprehensive experimental results on conditional image similarity tasks demonstrate that DIOR outperforms existing training-free baselines, including CLIP. Furthermore, DIOR achieves superior performance compared to methods that require additional training across multiple settings. |
| title | Training-free Conditional Image Embedding Framework Leveraging Large Vision Language Models |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.21860 |