Towards Fine-grained Visual Understanding and Visual Content Generation with Applications in Designing 

Avatar for Bingjie XU
Bingjie XU    
Assistant Professor

Read More 

Avatar for Indriyati ATMOSUKARTO
Indriyati ATMOSUKARTO    
Associate Professor

Read More 

Avatar for Daniel, Zhengkui WANG
Daniel, Zhengkui WANG    
Associate Professor

Read More 

Avatar for Jeffrey T.k.v. KOH
Jeffrey T.K.V. KOH    
Associate Professor

Read More 

This project explores the potential of AI to achieve detailed and nuanced visual comprehension, aimed at creating visual AIGC algorithm prototypes with fine-grained controllability, for creative content design, fashion design, and heritage visual storytelling.

Project Outcomes/Impact: 
•    Algorithm prototypes
•    UX applications
•    Publications
 

 

Diagram showing two technical overviews: (a) a multimodal system using an LLM and vision encoder for understanding and generation, and (b) a conceptual workflow for a diffusion model using source images, masks, and text prompts.