Cross-View Multi-Modal Segmentation @ Ego-Exo4D Challenges 2025
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866918047394037760 |
|---|---|
| author | Fu, Yuqian Wang, Runze Fu, Yanwei Paudel, Danda Pani Van Gool, Luc |
| author_facet | Fu, Yuqian Wang, Runze Fu, Yanwei Paudel, Danda Pani Van Gool, Luc |
| contents | In this report, we present a cross-view multi-modal object segmentation approach for the object correspondence task in the Ego-Exo4D Correspondence Challenges 2025. Given object queries from one perspective (e.g., ego view), the goal is to predict the corresponding object masks in another perspective (e.g., exo view). To tackle this task, we propose a multimodal condition fusion module that enhances object localization by leveraging both visual masks and textual descriptions as segmentation conditions. Furthermore, to address the visual domain gap between ego and exo views, we introduce a cross-view object alignment module that enforces object-level consistency across perspectives, thereby improving the model's robustness to viewpoint changes. Our proposed method ranked second on the leaderboard of the large-scale Ego-Exo4D object correspondence benchmark. Code will be made available at https://github.com/lovelyqian/ObjectRelator. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_05856 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Cross-View Multi-Modal Segmentation @ Ego-Exo4D Challenges 2025 Fu, Yuqian Wang, Runze Fu, Yanwei Paudel, Danda Pani Van Gool, Luc Computer Vision and Pattern Recognition Artificial Intelligence In this report, we present a cross-view multi-modal object segmentation approach for the object correspondence task in the Ego-Exo4D Correspondence Challenges 2025. Given object queries from one perspective (e.g., ego view), the goal is to predict the corresponding object masks in another perspective (e.g., exo view). To tackle this task, we propose a multimodal condition fusion module that enhances object localization by leveraging both visual masks and textual descriptions as segmentation conditions. Furthermore, to address the visual domain gap between ego and exo views, we introduce a cross-view object alignment module that enforces object-level consistency across perspectives, thereby improving the model's robustness to viewpoint changes. Our proposed method ranked second on the leaderboard of the large-scale Ego-Exo4D object correspondence benchmark. Code will be made available at https://github.com/lovelyqian/ObjectRelator. |
| title | Cross-View Multi-Modal Segmentation @ Ego-Exo4D Challenges 2025 |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2506.05856 |