Alex Kuefler
Papers
1
Total Citations
5
H-Index
1
About
Alex Kuefler is a roboticist whose research lies at the intersection of computer vision, manipulation, and self-supervised learning. His most cited work, "Depth by Poking: Learning to Estimate Depth from Self-Supervised Grasping" (2020, 5 citations), tackles a critical challenge in robotics: accurate depth estimation on reflective or transparent surfaces where traditional sensors like LiDAR and structured light fail. Kuefler’s key contribution is a self-supervised framework that trains a neural network to infer depth from RGB-D images by using physical interactions—specifically, a robot’s own grasping attempts—as supervisory signals. This approach not only bypasses the need for costly human-labeled data but also enables robots to learn robust depth perception directly from their environment. While his citation count is modest, the work is notable for its elegant, practical solution to a persistent problem in manipulation, demonstrating how active exploration can improve perceptual capabilities. Kuefler’s research is particularly relevant for students and researchers interested in closing the loop between perception and action, showing that robots can teach themselves to see better by simply trying to pick things up.
Research Focus
Key Achievements
Top Papers
- 1