Event-Based Visual Perception and Understanding

Event-based vision is emerging as a transformative paradigm for visual sensing, offering microsecond latency, high dynamic range, and ultra-low power consumption, while conventional image and video sensing remains the cornerstone of modern computer vision and multimedia analysis. Exploring both pure event-based vision and its combination with traditional image/video methods is crucial for tackling fast motion, high-contrast scenes, low-light environments, and resource-constrained platforms, despite the challenges posed by the asynchronous, sparse nature of event streams and the dense, synchronous structure of frames.

This workshop invites high-quality submissions on event-based vision and its integration with conventional vision, with a focus on novel models, algorithms, datasets, and systems for dynamic scene understanding, recognition, detection, tracking, motion analysis, and other visual tasks under realistic, high-speed, and high-dynamic-range conditions.