Zero-BEV: Zero-shot Projection of Any First-Person Modality to BEV Maps
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866914727273168896 |
|---|---|
| author | Monaci, Gianluca Antsfeld, Leonid Chidlovskii, Boris Wolf, Christian |
| author_facet | Monaci, Gianluca Antsfeld, Leonid Chidlovskii, Boris Wolf, Christian |
| contents | Bird's-eye view (BEV) maps are an important geometrically structured representation widely used in robotics, in particular self-driving vehicles and terrestrial robots. Existing algorithms either require depth information for the geometric projection, which is not always reliably available, or are trained end-to-end in a fully supervised way to map visual first-person observations to BEV representation, and are therefore restricted to the output modality they have been trained for. In contrast, we propose a new model capable of performing zero-shot projections of any modality available in a first person view to the corresponding BEV map. This is achieved by disentangling the geometric inverse perspective projection from the modality transformation, eg. RGB to occupancy. The method is general and we showcase experiments projecting to BEV three different modalities: semantic segmentation, motion vectors and object bounding boxes detected in first person. We experimentally show that the model outperforms competing methods, in particular the widely used baseline resorting to monocular depth estimation. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_13848 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Zero-BEV: Zero-shot Projection of Any First-Person Modality to BEV Maps Monaci, Gianluca Antsfeld, Leonid Chidlovskii, Boris Wolf, Christian Computer Vision and Pattern Recognition Robotics Bird's-eye view (BEV) maps are an important geometrically structured representation widely used in robotics, in particular self-driving vehicles and terrestrial robots. Existing algorithms either require depth information for the geometric projection, which is not always reliably available, or are trained end-to-end in a fully supervised way to map visual first-person observations to BEV representation, and are therefore restricted to the output modality they have been trained for. In contrast, we propose a new model capable of performing zero-shot projections of any modality available in a first person view to the corresponding BEV map. This is achieved by disentangling the geometric inverse perspective projection from the modality transformation, eg. RGB to occupancy. The method is general and we showcase experiments projecting to BEV three different modalities: semantic segmentation, motion vectors and object bounding boxes detected in first person. We experimentally show that the model outperforms competing methods, in particular the widely used baseline resorting to monocular depth estimation. |
| title | Zero-BEV: Zero-shot Projection of Any First-Person Modality to BEV Maps |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2402.13848 |