Gesture Detection in Digital Image Processing based on the Use of Convolutional Neuronal Networks

The paper presents empirical findings obtained from tests of an image detection system utilizing the YOLO network. Operating principles of the system are discussed, starting from simple convolutional networks, through R-CNN networks, and culminating in the use of the YOLO technology. The second part provides an examination of findings obtained from a system constructed on a custom dataset. Details are provided of the training process employed, together with the examination of accuracies for all pictures as well as for different classes. The main objective of the paper was to present the development of a system offering real-time gesture detection from camera feed images. Potential applications of the postulated gesture recognition system include toy control systems, photography, and broadly defined office assistance.