Towards Anytime Optical Flow Estimation with Event Cameras
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866908360934162432 |
|---|---|
| author | Ye, Yaozu Shi, Hao Yang, Kailun Wang, Ze Yin, Xiaoting Sun, Lei Wang, Yaonan Wang, Kaiwei |
| author_facet | Ye, Yaozu Shi, Hao Yang, Kailun Wang, Ze Yin, Xiaoting Sun, Lei Wang, Yaonan Wang, Kaiwei |
| contents | Event cameras respond to changes in log-brightness at the millisecond level, making them ideal for optical flow estimation. However, existing datasets from event cameras provide only low frame rate ground truth for optical flow, limiting the research potential of event-driven optical flow. To address this challenge, we introduce a low-latency event representation, Unified Voxel Grid, and propose EVA-Flow, an EVent-based Anytime Flow estimation network to produce high-frame-rate event optical flow with only low-frame-rate optical flow ground truth for supervision. Furthermore, we propose the Rectified Flow Warp Loss (RFWL) for the unsupervised assessment of intermediate optical flow. A comprehensive variety of experiments on MVSEC, DESC, and our EVA-FlowSet demonstrates that EVA-Flow achieves competitive performance, super-low-latency (5ms), time-dense motion estimation (200Hz), and strong generalization. Our code will be available at https://github.com/Yaozhuwa/EVA-Flow. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2307_05033 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Towards Anytime Optical Flow Estimation with Event Cameras Ye, Yaozu Shi, Hao Yang, Kailun Wang, Ze Yin, Xiaoting Sun, Lei Wang, Yaonan Wang, Kaiwei Computer Vision and Pattern Recognition Robotics Image and Video Processing Event cameras respond to changes in log-brightness at the millisecond level, making them ideal for optical flow estimation. However, existing datasets from event cameras provide only low frame rate ground truth for optical flow, limiting the research potential of event-driven optical flow. To address this challenge, we introduce a low-latency event representation, Unified Voxel Grid, and propose EVA-Flow, an EVent-based Anytime Flow estimation network to produce high-frame-rate event optical flow with only low-frame-rate optical flow ground truth for supervision. Furthermore, we propose the Rectified Flow Warp Loss (RFWL) for the unsupervised assessment of intermediate optical flow. A comprehensive variety of experiments on MVSEC, DESC, and our EVA-FlowSet demonstrates that EVA-Flow achieves competitive performance, super-low-latency (5ms), time-dense motion estimation (200Hz), and strong generalization. Our code will be available at https://github.com/Yaozhuwa/EVA-Flow. |
| title | Towards Anytime Optical Flow Estimation with Event Cameras |
| topic | Computer Vision and Pattern Recognition Robotics Image and Video Processing |
| url | https://arxiv.org/abs/2307.05033 |