Wide FoV and High Resolution Video Perception and Efficient Coding

Description

Objective: towards wide field of view (FoV) and high-resolution video collection and intelligent enhancement; anomaly event recognition, detection, and localization; and efficient video coding.Research Methodology: This research generally focuses on three aspects: theoretical models, method technology, and verification platforms. Aiming at complex public safety events, the core of the research generally focuses on super-resolution enhancement withlight-field images, abnormal video detection and tracking based on graph convolution networks and reinforcement learning, and efficient semantic coding of key events. Effects and significances: Monitoring of abnormal events in high-resolution video,perception and alarming. This research will have long-term effects on a series of fields such as public safety, transportation and criminal investigation.