Selected Bachelor’s theses

Automatic Detection of Metastases in Whole-Slide Lymph Node Images Using Deep Neural Networks

Author: Pavlína Koutecká, 2020
Supervisor: Jan Kybic

Digitisation of cancer recognition in histopathological images is researched topic in recent years, and automated computerised analysis based on deep neural networks has shown potential advantages as a diagnostic strategy. In this thesis, we develop a method for solving the task of automatic metastases detection in whole-slide lymph node images.

Description of epileptic network using distribution of interictal discharges

Author: Julie Barnová, 2020
Supervisor: Radek Janča

The irritative zone is a part of the concept of epileptic networks. It consists of subnetworks called clusters which can generate independent populations of interictal epileptiform discharges (IED). The aim of this thesis was to create and optimize an algorithm that would be able to identify independent sources of IED in intracranial EEG (iEEG) as well as demonstrate the relationship between their surgical removal and postoperative seizure freedom.

Selected Diploma theses

Localization and segmentation of in-vivo ultrasound carotid artery images

Author: Martin Kostelanský, 2021
Supervisor: Jan Kybic

This thesis is focused on the three separate image recognition tasks—classification, localization, and segmentation of the ultrasound images of the carotid artery with stenosis. The first problem was successfully solved by a ResNet50 CNN and a created dataset with 1,679 images.

Subtype Classification of Focal Cortical Dysplasia by Interictal Activity of Invasive EEG

Author: Laura Shala, 2020
Supervisor: Radek Janča

Focal cortical dysplasia (FCD) is a frequent cause of drug-resistant epilepsy. Determining the FCD subtype is essential in planning resection surgery in those patients. The aim of this work is to predict the FCD subtype based on the parameterization of the occurrence of interictal discharges IEDs in the preoperative iEEG recording.

Effective Connectivity Instability within Epileptogenic Network

Author: Lenka Svobodová, 2019
Supervisor: Radek Janča

One of the treatment options for pharmacoresistant epipsy is surgical removal of the epileptogenic zone. In the epileptogenic neural network we can identify nodes that are actively involved in the seizure genesis. These nodes show unstable effective connectivity with the rest of the network. The aim of this work is to identify unstable hubs, their localization in relation to clinical evaluation and the effect on surgical outcome.

Determination of Predictive Factors of Postoperative Atrial Fibrillation after Coronary Artery Bypass Grafting

Author: Kristýna Vieweghová, 2021
Supervisor: Radek Janča

This diploma thesis focuses on the statistical analysis of patient’s data after coronary artery bypass grafting (CABG). The main aim is to find predictors of the post-operative atrial fibrillation using multivariate statistical models. Theoretical part summarizes information about heart conduction system, evaluation of EKG and analysis of heart rate variability.

Secondary structure search in primary nucleic acid structures

Autor: Anh Vu Le, 2019
Vedoucí: Jiří Kléma

Structure of RNA molecules is often important for their function and regulation. Discovering a particular functional structure within a genome could then be viewed as discovering the associated function. This work proposes a computational pipeline, that searches the human genome and outputs potential IRES candidates

Neural Network Cascades to Incorporate Domain Knowledge for Hematopoietic Cell Classification

Autor: Jonas Daniel Nienhaus, 2020
Vedoucí: Philipp Gräbel, Jan Havlík

In this work, classification of haematopoietic cells is performed hierarchically, employing cascades of deep neural networks. Two strategies are defined to obtain combined predictions from the cascades, a probabilistic approach and a greedy, deterministic method.

Automatic intron detection in metagenomes using neural networks.

Autor: Martin Indra, 2021
Vedoucí: Jiří Kléma

Exact biological mechanisms of intron recognition and splicing are not fully known yet and their automated detection has remained unresolved. Detection and removal of introns from DNA sequences is important, e.g., for the identification of genes in metagenomes. Two neural network models were developed as part of this thesis. The models’ aim is the detection of intron starts and ends with the so-called donor and acceptor splice sites. The splice sites are later combined into candidate introns which are further filtered by a simple score-based overlap resolving algorithm.