Tolga Birdal

Stanford University, Imperial College London

Papers

Total Citations

H-Index

About

Tolga Birdal is a leading researcher in 3D computer vision and geometric deep learning, with a focus on primitive detection, object pose estimation, and shape understanding. His most-cited work, "From Planes to Corners," introduces a groundbreaking method for segmentation-free joint estimation of orthogonal planes, intersection lines, and corners in unorganized 3D point clouds. This unified approach enables robust scene exploration under orthogonality, with applications in semantic mapping and robotics, earning 11 citations. Birdal’s recent contributions include "NeRF-Feat," which leverages neural radiance fields for 6D object pose estimation from weakly labeled data, reducing reliance on costly CAD models or complex setups—a significant advance for robotic grasping and augmented reality. His work "Alignist" further advances orientation distribution estimation by fusing shape and correspondences, demonstrating his commitment to practical, data-efficient solutions. With a growing citation impact and a focus on bridging geometric primitives and learning-based methods, Birdal’s research is shaping the future of 3D scene understanding and autonomous systems.