Abstracts

Klasifikace nemoci na rentgenových snímcích pomocí umělé inteligence: Classification of thorax diseases on chest X-ray images using artificial intelligence

by Štěpán Pijáček

This thesis is researching workable solutions to the problem of classification of thorax disease on chest x-ray images using artificial intelligence. For a better understanding of the problem, the first chapters explain the basic convolutional neural network and its advantages and disadvantages. Based on these first explanations, two neural networks which are expanding on the concept of the convolutional neural network are chosen. Those are capsulated network and residual network both explained further in their respective sections with their advantages and disadvantages. Residual network and Capsulated network are implemented using programming language python and framework TensorFlow with Keras library, both with their respective chapters. At the end of this thesis, you can find results and conclusion.