A review: Study of various techniques of Hand Gestures Recognition

Abstract

With the development of Artificial Intelligence based augmented reality systems, pattern and gestures recognitions have evolved exponentially over time. They provide the flexibility to incorporate human gestures into digital world information that gives us another mode of human computer interaction. This paper presents an overview on several techniques opted by modern researchers to realize hand gestures recognition process. This whole process involves four steps comprising preprocessing of input frame, segmentation, features extraction and recognition process. In recent years, deep learning algorithms (Convolutional neural networks and Artificial neural networks) are the core attention to researchers because of their robust nature and ease of use. On the other hand, ordinary machine learning algorithms are less accurate and requires a lot of preprocessing beforehand but they are computationally faster than deep learning algorithms. This paper deals with a comparative analysis of statistical gestures recognition techniques and machine learning techniques.