Motivation While deep learning methods dominate image segmentation, they struggle with similar colors, complex textures, and require millions of annotated images. We propose a training-free approach that leverages shadow masks from photometric stereo data.
Sep 29, 2025
Overview Starting from a set of shadow masks, we use templates to extract light-to-shadow transitions. After combining these transitions into an edge strength and direction, we apply non-maximum suppression to obtain thin outlines.
Jul 25, 2024
Oct 1, 2022