Event-based Flow estimation using Events Demo

A real-world demo of work on "Adaptive-SpikeNet" showcased in the CVPR Event Vision Workshop. NRL presented a fully Spiking Neural Network (SNN) architecture with learnable neuronal dynamics to efficiently and effectively handle data from event-based cameras for estimating optical flow. Their models outperform current state-of-the-art ANN-based approaches while being energy-efficient and lightweight enough to run in real-time.