Towards Anytime Optical Flow Estimation with Event Cameras
This work addresses a bottleneck in event-based vision for applications requiring real-time motion estimation, though it is incremental in improving existing methods.
The paper tackles the problem of low-frame-rate ground truth in event camera optical flow estimation by introducing EVA-Flow, a network that produces high-frame-rate (200Hz) optical flow with low latency (5ms) and achieves competitive performance on benchmarks.
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.