Heart arrhythmias are a very common heart disease whose incidence is rising. This thesis is focused on the detection of premature ventricular contractions from 12-lead ECG records by means of deep learning. The location of these arrhythmias (key instances) in the record was found using a technique based on Multiple-Instance Learning. In the theoretical part of the thesis, basic electrophysiology of the heart and deep learning with a focus on the convolutional neural networks are described. Afterward, a program was created using the Python programming language, which contains a model based on the InceptionTime architecture, using which classification of the signals into the selected classes was performed. Grad-CAM was implemented to find locations of the key instances in the ECGs. The evaluation of the arrhythmia detection quality was done using the F1 score and the results were discussed at the end of the thesis.

This master's thesis describes basics principles of QRS complex detection. It is focused on QRS detection using zero crossing counts method. There are described princips and program realization of this method. The other part is focused on genetic optimalization algorithm. There are presented obtained optimalization results on standard CSE and MIT-BIH database. The quality of the detector is compared with other authors. The optimalized QRS detector achieves comparable results with other authors. The part of the thesis is graphical user interface which supply view on modified ECG signal and detection results.

Electrical impedance of biological matter is known as electrical bio-impedance or simply as bio-impedance. Bio-impedance devices are of great value for monitoring the pathological and physiological status of biological tissues in clinical and home environments. The technological progress in instrumentation has significantly contributed to the progress that has been observed during the last past decades in impedance spectroscopy and electrical impedance cardiograph. Although bio-impedance is not a physiological parameter, the method enables tissue characterization and functional monitoring and can contribute to the monitoring of the health status of a person. In this paper an inexpensive portable multi frequency impedance cardiograph device based on impedance spectroscopy technique has been developed. By means of this system the basic thoracic impedance range and the heart-action-caused changes of impedance can be measured and the hemodynamic parameters of the heart function can be determined. This system has small size and low current consumption. The impedance cardiograph signals of the electrodes configuration by Sramek, Penney and Qu in this work was measured; compared and summarized. The differences of the measuring method, the schematic circuit diagram, the measurement results and area of application between impedance cardiograph and impedance spectroscopy were discussed and compared. The performance of this sensor-system was evaluated.

This work deals with ECG signal measuring and methods of its processing. It compares some of the QRS detection methods and describes some of the testing databases. In this work a detector of QRS complex is realized, it is based on the approach of zero crossings. Next section makes combination of results from separate leads to one, which improves efficiency of detection. One section of this work deals with design and realization delination of ECG signal. In the last part outputs of this delineation are compared with the results of the other authors.

Heart dysfunction is a worldly widespread problem that currently is one of the leading causes of death. Studies indicate that many deaths related to cardiac dysfunction could have been prevented if discovered early. Contemporarily, ultrasound and electrocardiography are indispensable modalities for diagnostic purposes and analysis of cardiac function. The Ventricorder is an Ultra-Wideband radar sensor manufactured by the Norwegian company Novelda. Ventricorder has been shown to be able detect heart movements and breathing but its actual clinical value remains to be investigated. The Cardiac State Diagram (CSD) is a pre-clinical software tool for visualization of the heart's mechanical function. The CSD is confirmed by pilot studies to be able to constitute a basis for diagnosis and cardiac function assessment. Theoretically, the CSD is well suited to be used with the Ventricorder since the Ventricorder detects small changes over time and information about time events is all that is required for the creation of a CSD. Contemporarily, ultrasound tissue velocity imaging (TVI) is usually used for production of CSDs and in this master thesis we examined if the Ventricorder can be used to produce CSDs. This was done by mainly comparing velocity data from the Ventricorder with velocity data from temporally synchronized apical four-chamber images acquired with ultrasound TVI. The results indicate that there is an apparent correlation between these data sets and the Ventricorder should therefore be able to produce data that could constitute the basis for the production of a CSD. What remain now is to confirm these results statistically with a larger test group and to investigate whether all the time instants needed for the production of a CSD can be identified objectively. / Hjärtdysfunktion är ett värdsligt utbrett problem som ligger bakom många dödsfall varje år. Studier har visat att många dödsfall som är relaterade till hjärtdysfunktion hade kunnat förebyggas om de upptäckts i tid. För närvarande är bland annat ultraljud och EKG oumbärliga metoder för diagnostisering och analys av hjärtfunktion. Ventricorder är en typ av radarsensor som utnyttjar ett brett frekvensspektrum, så kallat Ultra Wideband, och är tillverkad av det norska företaget Novelda. Ventricorder har visat sig kunna detektera exempelvis hjärtrörelser och andning men dess kliniska värde har ännu inte undersökts. Cardiac State Diagram (CSD) är ett prekliniskt mjukvaruverktyg för att visualisera hjärtats mekaniska funktion och som har bekräftats genom pilotstudier att kunna användas som underlag för diagnostik och bedömning av hjärtats funktion. Teoretiskt sett är CSD väl lämpat för att användas med Ventricordern eftersom Ventricordern registrerar små rörelser över tid och just ändringar över tid är precis vad som behövs för att skapa ett CSD. I dagsläget används vanligen vävnadsdoppler (TVI) för produktion av CSD och i denna masteruppsats undersöktes huruvida Ventricorder kan användas för att producera CSD. Detta gjordes genom att jämföra mätdata från Ventricorder med temporalt synkroniserade apikala fyrkammar-bilder framställda med vävnadsdoppler. Resultaten indikerar att det finns en påtaglig korrelation mellan dessa data och att mätdatat från en Ventricorder således bör kunna användas för produktion av CSD. Det kvarstår att bekräfta dessa resultat statistiskt med en större testgrupp och att undersöka om samtliga tidsmarkörer som behövs för produktion av ett CSD kan identifieras objektivt.