Martin Danelljan

Papers

Total Citations

H-Index

About

Martin Danelljan is a prominent computer vision researcher whose work spans multiple object tracking, 3D scene reconstruction, and autonomous perception systems. He has made significant contributions to the challenge of understanding dynamic environments — a cornerstone of self-driving vehicles and robotic systems. His 2023 paper "OVTrack: Open-Vocabulary Multiple Object Tracking," which has garnered 58 citations, represents a landmark advance in breaking free from the rigid, category-limited frameworks that have long constrained traditional MOT benchmarks. By enabling tracking across an open vocabulary of object types, Danelljan's work dramatically broadens the applicability of tracking systems to real-world diversity. Complementing this, his work on "R3D3: Dense 3D Reconstruction of Dynamic Scenes from Multiple Cameras" (29 citations) tackles the formidable challenge of reconstructing complex, dynamic environments using cost-effective multi-camera setups — pushing the boundaries of what camera-only systems can achieve without expensive multi-modal hardware. Together, these contributions reflect Danelljan's overarching mission: making perception systems more generalizable, efficient, and deployable in the unpredictable complexity of the real world.