Tilemachos Aravanis received the best student paper award at the P13N - Personalization in Generative AI workshop at the International Conference on Computer Vision (ICCV) 2025 for his work "Only-style: Stylistic Consistency In Image Generation without Content Leakage". This work was the outcome of Tilemachos' undergraduate thesis at NTUA, Greece. Currently, he is a PhD student a VRG. 

Congratulations to Miroslav Purkrabek, Vladan Stojnic, Pavel Suma, Yaqing Ding, Giorgos Kordopatis-Zilos, Bill Psomas, Dmytro Mishkin, Tomas Vojir, and Giorgos Tolias.
 

Bill Psomas has been awarded the prestigious MSCA Postdoctoral Fellowship 2024 for his project RAVIOLI (Retrieval-Augmented VIsion-Language Models for Open-vocabulary LocalizatIon). Hosted at the Visual Recognition Group (VRG) at Czech Technical University in Prague (CTU) under the supervision of associate Prof. Giorgos Tolias, RAVIOLI aims to advance open-vocabulary segmentation by integrating retrieval-based predictions with vision-language models. The project seeks to enhance segmentation accuracy and adaptability across diverse environments and applications. As part of this research, Bill will also undertake a secondment at AImageLab, University of Modena and Reggio Emilia (UNIMORE), collaborating with Prof. Rita Cucchiara.

The Czech Science Foundation (GA ČR), the largest Czech grant funding provider for basic research, announced the results of the Standard grant projects for year 2025. The standard grand project are intended for applicants in all phases of their scientific career without age restrictions.The topic of a project is chosen by the applicant and the project is carried out by a single investigator or by research team from one or more institutions.

Researcher Tomáš Vojíř from Visual Recognition Group succeeded with his 3 year project  “Detection of anomalous or out-of-distribution inputs in deep neural network models”.

The "LUSt: Learning a Universal Similarity Function" project has been funded under Horizon Europe and will run for two years (2025-2026). This grant is supporting a postdoctoral fellowship at CTU in Prague with Giorgos Tolias as advisor and a collaboration with the University of Amsterdam and assistant professor Pascal Mettes. The project's objective is to perform research on similarity learning based on the similarities of the constituent parts of multimedia items. The ultimate goal is creating a foundational model capable of accommodating all modalities -- visual, audio, text -- and supporting a broad range of similarity types, including uni-modal, cross-modal, and multi-modal scenarios.

Yash Patel was announced as a winner of the Antonín Svoboda Award for the Best Ph.D. for Thesis defended in 2023. The Czech Society for Cybernetics and Informatics selects the best Ph.D. Thesis every year. The PhD thesis topic falls within the scope of ČSKI’s scientific activities (cybernetics, informatics, and related fields). Any Thesis defended in a particular year at a faculty located in the Czech Republic can be submitted.

The thesis title Neural Network Training and Non-Differentiable Objective Functions. Congratulations to Yash and his supervisor prof. Jiří Matas! More detailed article is available on the Faculty website here

The Dean’s Awards for prestigious dissertations were awarded on March 13th at the Scientific Council meeting. Researcher Yash Patel was awarded for his work Neural Network Training and Non-Differentiable Objective Functions under the supervision of Prof. Jiří Matas.