Showing cs.CVShow all
2 papers · 1 filter
cs.CV2026
Self-Verifier Capture in Unified Visual Generators: Diagnosing and Repairing Closed-Loop Self-Reward
Binyan Chen, Tong Pan
Image-space reward requires decoding and a separate judge; unified visual generators can instead reuse their understanding pathway for decode-free, self-supplied reward. Optimizing…
cs.CV2026
Keyframe-Anchored Identity Preservation for Sequential-Action Video Generation
Zhenjie Liu, Binyan Chen, Hao Chen +2
Identity-preserving text-to-video generation aims to synthesize a video that accurately follows a textual description while maintaining the recognizability of a user-specified subj…