Vladislav Golyanik

Max Planck Institute for Informatics, German Research Centre for Artificial Intelligence

Papers

Total Citations

204

H-Index

About

Vladislav Golyanik is a computer vision and graphics researcher whose work spans two interconnected frontiers: non-rigid 3D reconstruction and generative human motion synthesis. He has made significant contributions to the challenging problem of recovering three-dimensional structure from monocular video of deformable scenes — a fundamentally ill-posed problem that demands sophisticated mathematical and computational solutions. His 2017 work on dense batch non-rigid structure from motion demonstrated how convex relaxation and semidefinite programming could dramatically accelerate reconstruction pipelines, while his comprehensive 2023 state-of-the-art survey (36 citations) has become an essential reference for researchers entering the field. More recently, Golyanik has turned his attention to generative modeling for human motion, co-developing MoFusion, a denoising diffusion-based framework that elegantly resolves the long-standing tension between motion diversity and quality in conditional human motion synthesis. With 127 citations since 2023, MoFusion has rapidly established itself as a landmark contribution to the field. His work on HULC further demonstrates his commitment to physically grounded human motion capture, incorporating pose manifold sampling and dense contact guidance. Across his career, Golyanik has consistently pushed the boundaries of what is computationally achievable in understanding and synthesizing human movement and deformable scene geometry.