Non Line of Sight Reconstruction

Non-line-of-sight (NLOS) imaging aims to reconstruct hidden scenes by analyzing indirect light reflections, overcoming the limitations of direct line-of-sight imaging. Current research focuses on improving reconstruction accuracy and robustness, particularly in low signal-to-noise conditions, through the development of novel algorithms incorporating learned physical priors, differentiable rendering pipelines, and implicit neural representations like neural radiance fields (NeRFs). These advancements leverage both time-of-flight and frequency domain information, often employing optimization techniques like gradient descent and ADMM to solve the ill-posed inverse problem. The resulting improvements in NLOS imaging have significant implications for various fields, including robotics, surveillance, and medical imaging.