Ondrej Chum
Also known as Ondřej Chum
Ondrej Chum is a computer vision researcher at the Czech Technical University in Prague. After an MSc from Charles University in 2001, he completed a PhD at the Czech Technical University in 2005. He proposed PROSAC, a guided sampling method, and co-wrote the 2007 CVPR paper on object retrieval with large vocabularies, which received the Longuet-Higgins Prize in 2017. He is a full professor in the Department of Cybernetics and leads a team within the Visual Recognition Group.
- Works
- 131
- Citations
- 21,741
- h-index
- 41
Publication numbers from OpenAlex, as of 15 October 2026.
Timeline
-
2001
An MSc in Prague
He receives an MSc in computer science from Charles University in Prague in 2001.
-
2005
Guided sampling with PROSAC
At the Czech Technical University in Prague, he completes a PhD in 2005. He also proposes PROSAC, a method that guides the sampling procedure with prior information about the input data.
-
2006–2007
A postdoctoral researcher
From 2006 to 2007, he works as a postdoctoral researcher at the University of Oxford.
-
2007
Object retrieval with large vocabularies
With James Philbin, Michael Isard, Josef Sivic and Andrew Zisserman, he co-authors Object retrieval with large vocabularies and fast spatial matching, published at the Computer Vision and Pattern Recognition conference (CVPR) in 2007.
-
2017
A Longuet-Higgins Prize
In July 2017, at CVPR, the Longuet-Higgins Prize goes to the 2007 paper, which he shares with its co-authors.
-
2018
A Saburo Tsuji paper award
The Saburo Tsuji best paper award at ACCV 2018 goes to a paper on rectification from radially distorted scales, which he co-writes with James Pritts, Zuzana Kukelova and Viktor Larsson.
-
Since 2022
Full professor in Prague
He is one of 11 new full professors whom the Czech President announces, with diplomas due in January 2022. He leads a team within the Visual Recognition Group in the Department of Cybernetics.
Sources and credits
Photo: We have no published photo of Ondrej Chum yet. Send us a free-licensed photo.