Tomáš Karella

I am a Ph.D. student deeply passionate about computer vision and also a proud member of Fast Minds. My research focuses on the efficient training of deep learning models, leveraging a priori knowledge about the task at hand. Specifically, I am engrossed in topics like invariance, equivariance and robustness against expected transformations of the input data. A significant thrust of our work is to construct explainable networks that can ensure certain desired properties. Beyond the academic realm, I’m intrigued by the application of computer vision in nuclear energetics. This exploration is undertaken in collaboration with CVŘ.

Before my Ph.D. and even during my studies as a part-time endeavor, I worked at several companies. For a detailed account of my professional experience, please see my CV. On a personal note, I am an enthusiast of bouldering, ski-touring, and hiking. My hobbies even led me on two pilgrimages to Santiago de Compostela. Driven by my interest in deep learning, I initiated a reading group within our department, dedicated to the latest developments in deep learning.

Selected publications

H-NeXt: The next step towards roto-translation invariant networks

Tomáš Karella, Filip Šroubek, Jan Blažek, Jan Flusser, and Václav Košík

In 34th British Machine Vision Conference 2023, BMVC 2023, Aberdeen, UK, November 20-24, 2023, 2023

Abs Bib HTML PDF Code

The widespread popularity of equivariant networks underscores the significance of parameter efficient models and effective use of training data. At a time when robustness to unseen deformations is becoming increasingly important, we present H-NeXt, which bridges the gap between equivariance and invariance. H-NeXt is a parameter-efficient roto-translation invariant network that is trained without a single augmented image in the training set. Our network comprises three components: an equivariant backbone for learning roto-translation independent features, an invariant pooling layer for discarding roto-translation information, and a classification layer. H-NeXt outperforms the state of the art in classification on unaugmented training sets and augmented test sets of MNIST and CIFAR-10.
@inproceedings{karella2023, author = {Karella, Tomáš and Šroubek, Filip and Blažek, Jan and Flusser, Jan and Košík, Václav}, title = {H-NeXt: The next step towards roto-translation invariant networks}, booktitle = {34th British Machine Vision Conference 2023, {BMVC} 2023, Aberdeen, UK, November 20-24, 2023}, publisher = {BMVA}, year = {2023}, url = {https://papers.bmvc2023.org/0578.pdf}, }
Convolutional neural network exploiting pixel surroundings to reveal hidden features in artwork NIR reflectograms

Tomáš Karella, Jan Blažek, and Jana Striová

Journal of Cultural Heritage, 2022

Abs Bib PDF Code

Near-infrared reflectography (NIR) is a well-established non-invasive and non-contact imaging technique. The NIR methods are able to reveal concealed layers of artwork, such as a painter’s sketch or repainted canvas. The information obtained may be helpful to historians for studying artist technique, attributing an artwork reconstructing faded details. Our research presents the improved method previously developed that reveals the hidden features by removing the information content of the visible spectrum from NIR. Based on convolutional neural networks (CNN), our model estimates the transfer function from visible spectra to NIR, which is nonlinear and specific for painting materials. Its parameters are learnt for particular paintings on the subsamples randomly selected across the canvas, and the model is further utilised to enhance the whole artwork. In addition to the previously developed model, our algorithm exploits each pixel’s surroundings to estimate its NIR response. This leads to more precise results and increased robustness to various noises. We demonstrate higher accuracy than the previous method on the historical paintings mock-ups and higher performance on well-known artworks such as Madonna dei Fusi attributed to Leonardo da Vinci.
@article{karella2022, title = {Convolutional neural network exploiting pixel surroundings to reveal hidden features in artwork NIR reflectograms}, journal = {Journal of Cultural Heritage}, volume = {58}, pages = {186-198}, year = {2022}, issn = {1296-2074}, doi = {https://doi.org/10.1016/j.culher.2022.09.022}, author = {Karella, Tomáš and Blažek, Jan and Striová, Jana}, keywords = {Signal separation, Concealed features visualization, Artwork analysis, Infrared reflectography, Convolutional neural networks}, }

News

Dec 4, 2023	Just wrapped up my brief notes from BMVC 2023! 📝
Nov 20, 2023	🚀 Thrilled to announce my `first active participation` at #BMVC2023! Looking forward to sharing my work and networking with experts in the field. 😅 A bit nervous, it’s a big step for me! #ComputerVision
Oct 20, 2023	Just wrapped up my wonderful internship at Custom Ink 🎉 Loved enhancing their awesome images!🐙 #internship #CustomInk
Oct 10, 2023	Just had a great session sharing insights on AI wonders with the FTV Prima top management team! Exciting times ahead for media innovation! 🚀 #MediaTech #AINews
Jun 19, 2023	Thrilled to have attended #CVPR2023 in Vancouver - the forefront of Computer Vision innovation! 🤓🔍 #Science #ComputerVision