2022/03/30 by Daniel Lichy, Lichy, Daniel, Soumyadip Sengupta +3 · 1 citation
Computer Science · Environmental Science · #Advanced Vision and Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Optical measurement and interference techniques #Remote Sensing and LiDAR Applications
paper · pdf · doi:10.48550/arxiv.2203.16515
openalex publication_date 2022/03/30 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
We introduce the first end-to-end learning-based solution to near-field Photometric Stereo (PS), where the light sources are close to the object of interest. This setup is especially useful for reconstructing large immobile objects. Our method is fast, producing a mesh from 52 512×384 resolution images in about 1 second on a commodity GPU, thus potentially unlocking several AR/VR applications. Existing approaches rely on optimization coupled with a far-field PS network operating on pixels or small patches. Using optimization makes these approaches slow and memory intensive (requiring 17GB GPU and 27GB of CPU memory) while using only pixels or patches makes them highly susceptible to noise and calibration errors. To address these issues, we develop a recursive multi-resolution scheme to estimate surface normal and depth maps of the whole image at each step. The predicted depth map at each scale is then used to estimate `per-pixel lighting' for the next scale. This design makes our approach almost 45× faster and 2∘ more accurate (11.3∘ vs. 13.3∘ Mean Angular Error) than the state-of-the-art near-field PS reconstruction technique, which uses iterative optimization.