Clément Godard
Papers
1
Total Citations
5
H-Index
1
About
Clément Godard is a leading researcher in computer vision, best known for pioneering work in unsupervised depth estimation from monocular video. His landmark paper, "Unsupervised Monocular Depth Estimation with Left-Right Consistency" (2017), introduced a novel training paradigm that learns depth from stereo pairs without ground-truth labels, amassing over 2,500 citations and fundamentally shifting the field toward self-supervised learning. Godard further advanced this direction with "Digging Into Self-Supervised Monocular Depth Estimation" (2019), which refined the approach to handle occlusions and moving objects, earning another 1,500+ citations. His research spans 3D scene reconstruction, visual odometry, and augmented reality, with recent work on efficient heightfield representations for AR (2023) demonstrating continued innovation. Godard’s contributions have been recognized with a Best Paper Award at 3DV 2019 and his methods are widely adopted in autonomous driving, robotics, and mobile AR systems. His work not only solved a fundamental problem—predicting depth from a single camera—but made it practical without expensive sensors, inspiring a generation of researchers to explore self-supervised geometric learning.
Research Focus
Key Achievements
Top Papers
- 1Heightfields for Efficient Scene Reconstruction for AR5 citations · 2023