Shanghang Zhang

A new method improves how models perceive depth and spatial relationships in images.

SPHINX-V enhances AI's ability to interpret images through user interaction.

A new framework enhances AI's grasp of 3D spaces.

A novel method for creating detailed 3D images from single images using multiview diffusion.

CoCoGesture creates lifelike gestures that match spoken words, enhancing interaction.

A new model enhances the link between visual and language understanding.

MMTrail combines visual and audio descriptions for better video-language models.

FactorLLM improves efficiency in language models by reorganizing knowledge storage.

A new method enhances detail in image creation using regional prompts.

A novel approach enhances model learning from varied image data.

A new technique boosts image clarity in busy street environments.

Discover how ASGDiffusion changes high-resolution image generation.