报告题目：Mathematical Foundations of Non-Line-of-Sight Imaging

报 告 人：邱凌云（清华大学）

报告时间：2024年11月15日  16:00-17:00

报告地点：格物楼528

报告摘要：

Non-line-of-sight (NLOS) imaging has emerged as a promising technique for recovering obscured objects from scattered light. This research presents a unified mathematical framework for NLOS imaging in diverse scenarios, encompassing indoor and outdoor scenes, confocal and non-confocal settings, and significant background noise. The proposed framework incorporates sparsity, non-local self-similarity, and signal smoothness. By leveraging virtual confocal signals, our approach robustly reconstructs hidden objects' albedo and surface normal, even in the presence of high measurement noise and challenging spatial patterns of illumination and detection points. Synthetic and experimental results validate the effectiveness of our approach, surpassing state-of-the-art methods in terms of quantitative metrics and visual quality. This developed mathematical framework significantly advances NLOS imaging capabilities and holds, immense potential for applications across various fields.

报告人简介：

邱凌云，现任清华大学丘成桐数学科学中心助理教授，于2013年在美国普渡大学数学系获得博士学位。主要研究兴趣包括非线性反问题的分析与计算、最优输运理论、正则化方法、最优化问题的迭代算法以及深度学习在反问题上的应用。