Object identification from audio stream

This thesis deals with the classification of engine sound propulsion by motor vehicles. The purpose is to create a computerized system which is capable of identifying a specific motor vehicle engine sound. In thesis theory part is overviewing existing audio stream processing algorithms. Discussed about different machine learning algorithm. Describes other works directly related to thesis subject. In practical part is trying out one of already existing audio sound recognition programs. Later providing detailed description of developed deep learning model, tests results and conclusion of it