Ev2Gray: event-only intensity imaging – occlusion-referenced log-intensity reconstruction

Abstract This paper introduces Ev2Gray , a novel method for generating pseudo-grayscale images from pure event-camera data, addressing a core limitation of event-based vision: the absence of usable appearance information in static scenes. By sweeping a moving opaque stripe across the field of view, the approach imposes controlled occlusion and revelation, ensuring that every pixel briefly observes the same scene intensity. This establishes an intensity anchor across the sensor, enabling per-pixel values derived from polarity-resolved events that capture relative brightness structure without frames, training, or additional sensors. As a proof-of-concept, Ev2Gray demonstrates a transparent and reproducible pathway to recover large-scale structures, tone ramps, and fine details from event streams, making static content visually accessible. The simplicity and physical grounding of the method underscore its relevance for event-based vision pipelines that must infer the appearance from purely event streams.