PanoSAMic: Panoramic Image Segmentation from SAM Feature Encoding and Dual View Fusion
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866913058744434688 |
|---|---|
| author | Chamseddine, Mahdi Stricker, Didier Rambach, Jason |
| author_facet | Chamseddine, Mahdi Stricker, Didier Rambach, Jason |
| contents | Existing image foundation models are not optimized for spherical images having been trained primarily on perspective images. PanoSAMic integrates the pre-trained Segment Anything (SAM) encoder to make use of its extensive training and integrate it into a semantic segmentation model for panoramic images using multiple modalities. We modify the SAM encoder to output multi-stage features and introduce a novel spatio-modal fusion module that allows the model to select the relevant modalities and best features from each modality for different areas of the input. Furthermore, our semantic decoder uses spherical attention and dual view fusion to overcome the distortions and edge discontinuity often associated with panoramic images. PanoSAMic achieves state-of-the-art (SotA) results on Stanford2D3DS for RGB, RGB-D, and RGB-D-N modalities and on Matterport3D for RGB and RGB-D modalities. https://github.com/dfki-av/PanoSAMic |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_07447 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | PanoSAMic: Panoramic Image Segmentation from SAM Feature Encoding and Dual View Fusion Chamseddine, Mahdi Stricker, Didier Rambach, Jason Computer Vision and Pattern Recognition Existing image foundation models are not optimized for spherical images having been trained primarily on perspective images. PanoSAMic integrates the pre-trained Segment Anything (SAM) encoder to make use of its extensive training and integrate it into a semantic segmentation model for panoramic images using multiple modalities. We modify the SAM encoder to output multi-stage features and introduce a novel spatio-modal fusion module that allows the model to select the relevant modalities and best features from each modality for different areas of the input. Furthermore, our semantic decoder uses spherical attention and dual view fusion to overcome the distortions and edge discontinuity often associated with panoramic images. PanoSAMic achieves state-of-the-art (SotA) results on Stanford2D3DS for RGB, RGB-D, and RGB-D-N modalities and on Matterport3D for RGB and RGB-D modalities. https://github.com/dfki-av/PanoSAMic |
| title | PanoSAMic: Panoramic Image Segmentation from SAM Feature Encoding and Dual View Fusion |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2601.07447 |