Sparse Views, Near Light: A Practical Paradigm for Uncalibrated Point-light Photometric Stereo
This work addresses the impracticality of traditional photometric stereo methods, which require dark rooms and lab conditions, by enabling easy data capture outside the lab for applications needing accurate 3D reconstruction.
The paper tackles the problem of high-quality 3D reconstruction from sparse viewpoints in real-world environments by proposing an end-to-end uncalibrated multi-view photometric stereo framework, which outperforms existing baselines by a large margin in sparse-view regimes.
Neural approaches have shown a significant progress on camera-based reconstruction. But they require either a fairly dense sampling of the viewing sphere, or pre-training on an existing dataset, thereby limiting their generalizability. In contrast, photometric stereo (PS) approaches have shown great potential for achieving high-quality reconstruction under sparse viewpoints. Yet, they are impractical because they typically require tedious laboratory conditions, are restricted to dark rooms, and often multi-staged, making them subject to accumulated errors. To address these shortcomings, we propose an end-to-end uncalibrated multi-view PS framework for reconstructing high-resolution shapes acquired from sparse viewpoints in a real-world environment. We relax the dark room assumption, and allow a combination of static ambient lighting and dynamic near LED lighting, thereby enabling easy data capture outside the lab. Experimental validation confirms that it outperforms existing baseline approaches in the regime of sparse viewpoints by a large margin. This allows to bring high-accuracy 3D reconstruction from the dark room to the real world, while maintaining a reasonable data capture complexity.