Enregistré dans:
| Auteurs principaux: | , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2025
|
| Sujets: | |
| Accès en ligne: | https://arxiv.org/abs/2512.04970 |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912749480574976 |
|---|---|
| author | Pogorelyuk, Leonid Bracher, Niels Verkleeren, Aaron Kühmichel, Lars Radev, Stefan T. |
| author_facet | Pogorelyuk, Leonid Bracher, Niels Verkleeren, Aaron Kühmichel, Lars Radev, Stefan T. |
| contents | We pilot a family of stable contrastive losses for learning pixel-level representations that jointly capture semantic and geometric information. Our approach maps each pixel of an image to an overcomplete descriptor that is both view-invariant and semantically meaningful. It enables precise point-correspondence across images without requiring momentum-based teacher-student training. Two experiments in synthetic 2D and 3D environments demonstrate the properties of our loss and the resulting overcomplete representations. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2512_04970 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Stable Single-Pixel Contrastive Learning for Semantic and Geometric Tasks Pogorelyuk, Leonid Bracher, Niels Verkleeren, Aaron Kühmichel, Lars Radev, Stefan T. Computer Vision and Pattern Recognition We pilot a family of stable contrastive losses for learning pixel-level representations that jointly capture semantic and geometric information. Our approach maps each pixel of an image to an overcomplete descriptor that is both view-invariant and semantically meaningful. It enables precise point-correspondence across images without requiring momentum-based teacher-student training. Two experiments in synthetic 2D and 3D environments demonstrate the properties of our loss and the resulting overcomplete representations. |
| title | Stable Single-Pixel Contrastive Learning for Semantic and Geometric Tasks |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2512.04970 |