A Shadow Based Low-Cost Hand Movement Recognition System for Human Computer Interaction

Abstract

TIn this paper, we propose and implement a real-time hand motion detection system which uses shadows projected by hand and detected by low cost Light Dependent Resistor (LDR). The main advantage of the shadow approach is real-time recognition and its application in Post COVID 19 world where contactless interactions are required. Our proposed approach uses 2D hand motion shadows to determine sequence of light blocking the LDRs and thus making the system less complex and less compute intensive compared to 3D image recognition based systems. The accuracy of the proposed system is tested under various lighting conditions and 82-95\\% accuracy is observed under normal lighting conditions.