Efficient Hand Gesture Recognition Using Modified Extrusion Method based on Augmented Reality

Abstract

Recognition of hand gesture with efficient methodology is an unsolved research problem in order to apply in education sector specially for improving learning process. In this context, mathematics is in great consideration by the researchers to make the process of learning easier using technology especially for the learners who are in the preliminary stage. Previously, research involved in improving learning process of mathematics using gesture recognition could not progress in large extent due to lack of comprehensive experimental validation and high computational complexity in lieu with less cost effective, deformation of hands, color of skin or unusual gesture. This research proposes modified extrusion method for hand gesture recognition which converts 2D shape to 3D shape in order to control these augmented geometrical shapes by gesture. In addition, proposed modified extrusion method is validated with standard performance metrics, i.e. accuracy, error rate and processing speed based on two perspectives, i.e. tracking and non-tracking of hand. Proposed method achieved accuracy rate of 95% in lieu with 10% error rate which reveals better performance comparing with previous research methods. Thus, experimental results reveal the efficiency of the proposed methodology.