Metric Depth

Metric depth estimation aims to recover accurate, real-world distances from images, a challenging problem due to the inherent scale ambiguity in single-view projections. Current research focuses on improving the generalization of monocular depth estimation models across diverse scenes and camera parameters, employing techniques like incorporating language descriptions, robot kinematics, ground constraints, and diffusion models to refine relative depth maps into metric scales. These advancements are crucial for robotics, augmented reality, and 3D scene reconstruction, enabling more robust and accurate applications in these fields.