|
Implementation of new method to machine learning model for epileptogenic zone localization in pharmacoresistant epilepsy patients
Pivnička, Martin ; Mívalt, Filip (referee) ; Filipenská, Marina (advisor)
The bachelor thesis describes the issue of the epileptogenic tissue localization considering pacients with drug-resistant epilepsy. The first half of the theoretical part discusses the matter of epilepsy and its treatment. It describes the principle of electroencephalographic measurement and its contribution to epileptology as well as multiple foci localization approaches. The second theoretical part shows machine learning basics and its use for epilepsy treatment. The practical part starts with the description of steps needed to create the gamma method. It continues with the statistical analysis of the method. This analysis contains both gamma method alone and as a part of existing machine learning algorithm. It has been shown that the gamma method is a valuable specific parameter for localizing epileptic foci. Its addition to the machine learning model did not lead to a significant improvement in the performance of the model.
|
|
Automatic -wave detection in 12-lead ECG
Khunová, Martina ; Filipenská, Marina (referee) ; Ředina, Richard (advisor)
This bachelor thesis deals with the automatic detection of delta waves from the 12-lead ECG in Matlab. In the theoretical part, the anatomy and physiology of the heart is briefly described, the reader gets familiar with Wolff-Parkinson-White syndrome, and through the manifestations of delta waves on the electrocardiogram we come to the description of linear filters and detection of the QRS complex based on the envelope. In the first part of practical part, a QRS complex detector is constructed, which is followed by a delta wave detector. The detection of the delta wave is based on the measurement of the duration of the peak and its derivation. The detector was tested on a database which data comes from pediatric patients.
|
|
Advanced sleep quality estimation
Benáček, Petr ; Ředina, Richard (referee) ; Filipenská, Marina (advisor)
This thesis deals with the assessment of sleep quality using modern deep learning methods. The thesis describes metrics for automatic classification of sleep stages. A selected database of sleep data is discussed. Due to the low number of data in the wakefulness phase, different methods of data augmentation are described and implemented. Models based on 1D convolutional networks are the basis for the classification. As a result, models for binary classification and classification of 3 and 4 sleep phases are prepared. Finally, sleep quality metrics are calculated using these models and the results are compared with the literature.
|
|
Advanced sleep scoring
Dokoupilová, Daniela ; Novotná, Petra (referee) ; Filipenská, Marina (advisor)
This diploma thesis focuses on classification of sleep stages using a smart watch. Two signals were used – heart rate and acceleration. A model called TinySleepNet composed of convolutional neural network and LSTM was chosen for this task. The model was first trained for the classification of five sleep stages using only heart rate, achieving F1 score of 49%. Acceleration was converted into an SVM vector, on which the second model was trained. Due to the lack of information in the SVM vector, the model was trained only for binary classification of wake/sleep, achieving F1 score of 62.3%. Both SVM and heart rate were combined in the last model. The classification of heart rate and SVM vector into five sleep stages achieved F1 score of 51%. The calculated parameters evaluating sleep quality were then compared with data evaluated by a sleep expert.
|
|
Health assessment using smart devices
Vargová, Enikö ; Filipenská, Marina (referee) ; Němcová, Andrea (advisor)
This thesis deals with the possibilities of non-invasive determination of blood glucose from photoplethysmographic signals. Elevated blood sugar is often associated with disease called diabetes mellitus. Diabetes is one of the world’s major chronic diseases. Untreated diabetes is often a cause of death. The aim of the work is to propose methods for glycemic classification and prediction. Two datasets have been created by recording the PPG signals using two smart devices (a smart wristband and a smartphone), along with their blood glucose levels measured in an invasive way. The PPG signals were preprocessed, and suitable features were extracted from them. Various machine-learning models for glycemic classification and prediction were created.
|