Research of convolutional neural networks for games using visual recognition

In this thesis is analyzed workflow of convolutional neural networks and their adaptation in first-person shooter game Counter-Strike: Global Offensive by identifying and classifying players encountered during the game. Thesis describes the basic operations, techniques that are commonly used to apply in the image classification using convolutional neural networks. Also, this paper describes architectures used in convolutional neural networks. The paper reviews most popular these days existing frameworks and common problems faced for solving this problem. The paper compares results obtained from different architectures used to develop object detection models