Motion Compensated Compression for Event-Based Cameras

Event-based cameras detect significant log-luminance changes at each pixel asynchronously and report these events in real-time. A changing scene can generate many events in a short time, resulting in the need for data compression. In this paper, we propose a lossless Motion Compensated Compression algorithm based on Optical Flow (MCCOF) for event cameras. We show that the MCCOF achieves better performance on most datasets compared to the previous spike coding algorithm. Using a real-time event-based optical flow algorithm for motion compensation, the algorithm does not significantly increase the computational time for compression and decompression compared to previous algorithms.