Research

My Ph.D. research focuses on algorithms for visual amodal perception, which enables vision models to infer and understand parts of the visual world that are not directly visible. Moving forward, I'm interested in exploring its role in spatial intelligence, world models, and robotics, particularly in enabling AI systems to reason about hidden scene structure, understand physical environments, and act effectively under partial observation.

I also participate in research on medical imaging, AI for energy and environment, robotics and 3D vision, and virtual try-on.