Identifiable Object Representations under Spatial Ambiguities
Fuente:
arXiv
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866909643719049216 |
|---|---|
| author | Kori, Avinash Toni, Francesca Glocker, Ben |
| author_facet | Kori, Avinash Toni, Francesca Glocker, Ben |
| contents | Modular object-centric representations are essential for *human-like reasoning* but are challenging to obtain under spatial ambiguities, *e.g. due to occlusions and view ambiguities*. However, addressing challenges presents both theoretical and practical difficulties. We introduce a novel multi-view probabilistic approach that aggregates view-specific slots to capture *invariant content* information while simultaneously learning disentangled global *viewpoint-level* information. Unlike prior single-view methods, our approach resolves spatial ambiguities, provides theoretical guarantees for identifiability, and requires *no viewpoint annotations*. Extensive experiments on standard benchmarks and novel complex datasets validate our method's robustness and scalability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_07806 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Identifiable Object Representations under Spatial Ambiguities Kori, Avinash Toni, Francesca Glocker, Ben Machine Learning Computer Vision and Pattern Recognition Modular object-centric representations are essential for *human-like reasoning* but are challenging to obtain under spatial ambiguities, *e.g. due to occlusions and view ambiguities*. However, addressing challenges presents both theoretical and practical difficulties. We introduce a novel multi-view probabilistic approach that aggregates view-specific slots to capture *invariant content* information while simultaneously learning disentangled global *viewpoint-level* information. Unlike prior single-view methods, our approach resolves spatial ambiguities, provides theoretical guarantees for identifiability, and requires *no viewpoint annotations*. Extensive experiments on standard benchmarks and novel complex datasets validate our method's robustness and scalability. |
| title | Identifiable Object Representations under Spatial Ambiguities |
| topic | Machine Learning Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.07806 |