Event Transformer
This work addresses the challenge of efficiently processing event camera data for computer vision tasks, offering a method that preserves temporal granularity without specialized hardware, though it appears incremental in its approach.
The paper tackled the problem of processing event camera data by introducing a novel token-based event representation and a Three-way Attention mechanism in the Event Transformer Block, achieving competitive performance in object classification and optical flow estimation with minimal computational resources on standard devices.
The event camera's low power consumption and ability to capture microsecond brightness changes make it attractive for various computer vision tasks. Existing event representation methods typically convert events into frames, voxel grids, or spikes for deep neural networks (DNNs). However, these approaches often sacrifice temporal granularity or require specialized devices for processing. This work introduces a novel token-based event representation, where each event is considered a fundamental processing unit termed an event-token. This approach preserves the sequence's intricate spatiotemporal attributes at the event level. Moreover, we propose a Three-way Attention mechanism in the Event Transformer Block (ETB) to collaboratively construct temporal and spatial correlations between events. We compare our proposed token-based event representation extensively with other prevalent methods for object classification and optical flow estimation. The experimental results showcase its competitive performance while demanding minimal computational resources on standard devices. Our code is publicly accessible at \url{https://github.com/NJUVISION/EventTransformer}.