Camera Projection

Camera projection research focuses on accurately modeling how 3D scenes are mapped onto 2D images, crucial for tasks like 3D reconstruction, augmented reality, and autonomous navigation. Current research emphasizes robust methods for handling diverse camera types and challenging scenarios, including developing novel neural light field models for panoramic image stitching, improving monocular visual odometry initialization, and employing geometric projectors for efficient optimization in robot navigation. These advancements are driving improvements in applications ranging from mobile phone panoramas to autonomous vehicle perception and 3D modeling from images.

Papers