Improvement of Vision-Based Hand Gesture Recognition System with Distance Range

Hand gestures are widely explored for the human action interface, reducing the complexity of interaction between humans and computers. Hand gesture-based interaction for computer applications enables touchless operation without the need for other devices such as a keyboard and mouse. It also guarantees sterility and safer interaction for multiple users. The issue of HGR has a low range of detection. This study aims to improve the distance range for HGR from the user to the computer camera using the zoom algorithm that is processed by OpenCV and the Mediapipe library. To verify the proposed system, a Graphical User Interface (GUI) for four applications is developed in HGR mode. The GUI includes the development of cursor mouse event commands and a virtual keyboard display. The proposed method was then tested for some time, and the result indicates that the maximum range of detection increased with the accuracy of the HGR. The proposed method meets the expected outcome and can be used for applications such as slide presentations and watching video applications.