Noah Snavely

Papers

Total Citations

H-Index

About

Noah Snavely is a leading figure in computer vision and graphics, renowned for pioneering large-scale 3D reconstruction from Internet imagery. His foundational work on "Photo Tourism" and "Building Rome in a Day" revolutionized the field by demonstrating how unordered photo collections could be automatically aligned to produce navigable 3D point clouds, earning over 5,000 citations collectively. Snavely's research spans structure from motion, scene understanding, and neural rendering, with major contributions including the development of SIFT-based matching pipelines and incremental SfM algorithms that became standard tools. His paper "An Analysis of SVD for Deep Rotation Estimation" (2020, 32 citations) provides theoretical grounding for rotation estimation in deep learning, while "Stereo4D" (2025) pushes boundaries in dynamic 3D scene understanding from Internet video. Snavely's work has been recognized with multiple best paper awards at CVPR and SIGGRAPH, and he currently leads the Cornell Tech Computer Vision Lab. His research has directly enabled applications in autonomous navigation, augmented reality, and cultural heritage preservation, making him a pivotal figure in democratizing 3D reconstruction.