Hand Detection and Body Language Recognition Using YOLO

Abstract

Neural Networks play an important role in real-time object detection. Several types of networks are being developed in order to perform such detections at a faster pace. One such neural network that can prove useful is the YOLO network. Built to perform real-time detection, YOLO offers great speeds for simple detections. The goal of our research is to see how YOLO would work with body language. Would it be fast enough? And how accurate would it be? Compared to other forms of object detection, body-language detection is more vague. There are several factors to be accounted for. This is why we first begin by talking about hand recognition and gesture recognition, and then move onto body language. This research aims at understanding how YOLO would perform when subject to several tests by using its implementations, building datasets, training and testing the models to see whether it is successful in detecting hand gestures and body language.