Mike Zheng Shou

New methods improve multimodal models by adding questioning and evaluation tasks.

WMAdapter simplifies watermarking for AI-generated images while ensuring quality and effectiveness.

Examining vulnerabilities in digital watermarking methods and their implications for media protection.

A new approach enhances video data handling while maintaining performance in real-time applications.

Learn how machine learning models can improve when facing new and unseen data.

VideoLISA uses language to segment and track objects in videos effectively.

A new method improves realism in human image animations for various applications.

Change how you see videos with ReCapture's innovative angle shifting technology.

A new dataset aims to improve long video storytelling and character consistency.

Factorized Quantization improves image generation through efficient token management.

Anti-Reference safeguards your images from misuse and manipulation.

New method safeguards personal images from misuse and identity theft.

Discover how DiffSim transforms image comparison with advanced techniques.