Study of deep learning techniques for human gesture classification

IRIIn this project we will propose a solution for human gesture classification using Deep Learning models, meant to be useful for non-verbal comunication between humans and robots. We have created a human body gesture dataset using a MediaPipe pose tracking solution in order to train the main model of this work. The MediaPipe solution uses Convolutional Neural Networks for body joints extraction and our model implements Deep Neural Networks for classification. This is meant to be an introductory work to the area of Deep Learning