The application of deep learning methods in the identification of abnormalities on the CT images of the brain using gpus

Burkhonov Ravshan Abdujabborovich

The aim of the study is to apply the methods of deep learning in the detection of pathology on the brain tomograms using graphic processors. The objectives of the study are the analysis of deep learning methods, approaches to the detection of pathology on the brain tomograms, the use of the CUDA library in the detection of pathologies. Hypothesis of the study: using graphics processors will increase the speed and quality of pathology detection on the brain tomograms. Research methods: analysis, synthesis, comparison and analogies. Achieved results: the obtained models are convolutional neural networks for definition of the norm and different pathologies on CT images of the brain.

Keywords:deep learning, neural networks, machine learning, brain tomograms, brain pathologies, recognition, CUDA libraries