FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866910977259208704 |
|---|---|
| author | Chou, Gene Xian, Wenqi Yang, Guandao Abdelfattah, Mohamed Hariharan, Bharath Snavely, Noah Yu, Ning Debevec, Paul |
| author_facet | Chou, Gene Xian, Wenqi Yang, Guandao Abdelfattah, Mohamed Hariharan, Bharath Snavely, Noah Yu, Ning Debevec, Paul |
| contents | A versatile video depth estimation model should (1) be accurate and consistent across frames, (2) produce high-resolution depth maps, and (3) support real-time streaming. We propose FlashDepth, a method that satisfies all three requirements, performing depth estimation on a 2044x1148 streaming video at 24 FPS. We show that, with careful modifications to pretrained single-image depth models, these capabilities are enabled with relatively little data and training. We evaluate our approach across multiple unseen datasets against state-of-the-art depth models, and find that ours outperforms them in terms of boundary sharpness and speed by a significant margin, while maintaining competitive accuracy. We hope our model will enable various applications that require high-resolution depth, such as video editing, and online decision-making, such as robotics. We release all code and model weights at https://github.com/Eyeline-Research/FlashDepth |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2504_07093 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution Chou, Gene Xian, Wenqi Yang, Guandao Abdelfattah, Mohamed Hariharan, Bharath Snavely, Noah Yu, Ning Debevec, Paul Computer Vision and Pattern Recognition A versatile video depth estimation model should (1) be accurate and consistent across frames, (2) produce high-resolution depth maps, and (3) support real-time streaming. We propose FlashDepth, a method that satisfies all three requirements, performing depth estimation on a 2044x1148 streaming video at 24 FPS. We show that, with careful modifications to pretrained single-image depth models, these capabilities are enabled with relatively little data and training. We evaluate our approach across multiple unseen datasets against state-of-the-art depth models, and find that ours outperforms them in terms of boundary sharpness and speed by a significant margin, while maintaining competitive accuracy. We hope our model will enable various applications that require high-resolution depth, such as video editing, and online decision-making, such as robotics. We release all code and model weights at https://github.com/Eyeline-Research/FlashDepth |
| title | FlashDepth: Real-time Streaming Video Depth Estimation at 2K Resolution |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2504.07093 |