Face and Gesture Based Human - Computer Interaction System

DOI

No DOI

Conference

Innovations in Computing, Materials and Communication Technologies

Organised By

IES College of Engineering

Conference Date

15 May 2023

Pages

69-76

Article Type

Research Article

Check for updates

No DOI available for cross-reference

Computers and people interact in a number of ways, and these interactions are made possible by the interface between the two. To make interactions between people and computers as simple and efficient several methods has been proposed. The vision-based systems enabling facial and hand gesture identification are the most promising ones for contactless human-computer interaction (HCI). This project proposed a human-computer interaction (HCI) system by combining a face and gesture recognition system. Face recognition based authentication system is more popular than any other biometric features like fingerprint and eye iris recognition. The face recognition system performs identifying and verifying a person in front of the camera. If the image matches with the photo in the database, the face is labelled with the person's name and a successful login is done within system. If it does not match, the image displays unknown and an unauthorized access is detected. Here face detection and recognition is implemented using HAAR-cascade classifiers and LBPH recognizers. The hand gesture recognition system recognises the gesture and links it with a variety of actions, including launching programmes like windows media player, MS word, PowerPoint, Notepad, screenshot capturing, opening Google and some of the mouse control actions. Here hand tracking performs with the help of Media Pipe library provided by Google which is open-source and it is a well-trained model to achieve high performance. The Media Pipe library can be used to analyse hand gestures using a variety of technologies. Media Pipe hands library will employ two models. 1) A palm detector model that generates a bounding box of hand and 2) A hand landmark model that predicts the hand skeleton. Gestures can be used by users to effortlessly communicate with computers that have RGB cameras.