A low-cost hand gesture and face recognition system for non-contact applications

Developments in noncontact technologies have attracted much attention for diverse and innovative applications. The desired functions can be activated by contactless interactions, such as hand gestures and face recognition. A contactless control system would need to be inexpensive, compact, easy to use, and energy efficient for widespread adoption. The system can be efficiently implemented thanks to advances in deep learning (DL) and computer vision (CV). In this study, a proof-of-concept prototype of the designed hand gesture recognition system is shown to control devices by successfully recognizing gestures that represent from digits zero to nine. An input image of 32 × 32 pixels is sufficient to extract the background and finger and requires less than 10% of the computational cost of typical high-resolution image convolution. As a result, it demonstrates cost-effective potential. In addition, a noncontact user-friendly interface that allows users to gain access and store their information without physical interaction is discussed. The proposed contactless access control strategy randomly instructs users to perform specific face actions and evaluates these actions to prevent spoofing and to further enhance security. The integration of a noncontact method with recognition technology enhances the opportunity to contribute to wide practical applications in the future.