Recent Advances in Vision Foundation Models

CVPR 2025 Tutorial on "Recent Advances in Vision Foundation Models"

We present our CVPR tutorial proposal on Recent Advances in Vision Foundation Models, a topic that has garnered significant attention from the computer vision community. Our tutorial will cover the most advanced directions in designing and developing vision foundation models, including the state-of-the-art approaches and principles in (i) learning vision foundation models for multimodal understanding and generation, (ii) scaling test-time compute and enabling the self-training of foundation models to improve themselves on reasoning and perception, and (iii) physical and virtual agents based on vision foundation models that can take actions for robotics and in virtual environments.