G. Tolias - Γ. Τόλιας

Giorgos Tolias

I am an associate professor at CTU in Prague and a member of the Visual Recognition Group (VRG). I am currently leading a small team within VRG.

From 2014 to mid-2015 I was a postdoc at the LinkMedia team of Inria Rennes, while I obtained my Ph.D from the NTU of Athens in 2013, where I worked with the Image and Video Analysis Team (IVA). In the past, I have worked closely with Ondrej Chum, Herve Jegou, and Yannis Avrithis.

My research is on computer vision. I enjoy working on learning visual representation, especially in the limited supervision regime. A special focus is on recognition problems (i) for large category sets, (ii) with a fine granularity of labels, or (iii) under a distribution shift.


PhD students


Oct2023:Paper on active learning accepted at WACV 2024.

Oct2023:Paper on test-time training for one-shot video object segmentation accepted at NeurIPS 2023.

June2023:Tutorial talk at CVPR 2023 on Large-scale visual localization.

Oct2022:Co-organizing the Instance-level Recognition Workshop and the Sketching for Human Expressivity Workshop at ECCV 2022.

Aug2022:Paper on resolution-wise asymmetric metric learning accepted at WACV 2023.

Jun2022:Paper on large-scale sketch recognition accepted at ICPR 2022.

Mar2022:Paper on recall loss and similarity mixup accepted at CVPR 2022.

Nov2021:Paper on artwork recognition accepted at NeurIPS 2021.

Oct2021:Speaker at the Large-scale Visual Localization tutorial at ICCV 2021. See video here.

Oct2021:Co-organizing the Instance-level Recognition Workshop at ICCV 2021.

Jun2021:Co-organizing the Image Similarity Challenge.

Nov2020:I got awarded a Junior Star grant from the Czech Science foundation.

Oct2020: New but old: Python implementation of ASMK: retrieval approach with local descriptors

Jul2021:Two papers accepted at ECCV 2020: local descriptors for instance recognition (oral) and learning with few clean and many noisy examples.

toliageo -- a t -- fel.cvut.cz
Karlovo namesti 13, 121 35 Praha 2, Czech Republic
Office: G10 (building G, room 10),