We plan to design cutting-edge algorithms to enable fast and generalizable 3D reconstruction, novel view synthesis, and 3D-aware image synthesis by combining advances in computer vision and computer graphics. This project includes three main parts: (1) Geometric modeling and reconstruction are fundamental to computer vision and graphics. The classical point graph geometric modeling and feature matching-based reconstruction methods have long served as the workhorse of 3D digital applications. (2) Realistic appearance rendering. Reconstructing spatial-varying reflectance or radiance field in public spaces has increased dramatically in recent years. Our focus is to develop lightweight, fast and generalizable solutions by coupling inverse physical-based rendering with new computer vision algorithms. (3) Controllable image synthesis conditioning on labels, for instance, 2D semantic segmentation map, text, camera poses, geometric, material, and lighting parameters, etc.