Surface Normal Estimation
Overview
Surface normal estimation predicts the orientation of surfaces at image pixels, supporting 3D reconstruction, robotics, inspection, and tactile perception. Approaches include monocular neural networks and diffusion models, photometric stereo with controlled lighting, shape-from-polarization, and geometric cues from stereo or video. Recent methods combine learned priors with physical measurements and consistency constraints to handle reflective, transparent, underwater, or sparsely observed surfaces. A central challenge is generalizing across scenes and sensors while preserving fine detail and geometric consistency despite ambiguous appearance, limited ground truth, and calibration demands.
Latest papers
26 papers4 papers in the last four weeks, up 33% on the four weeks before. 0.0% of all new papers.