Towards real-time visual biometric authentication using human face for healthcare telepresence mobile robots

• Main Authors: Muralindran Mariappan, Manimehala Nadarajan, Rosalyn R Porle, Norfarariyanti Parimon
• Format: Article
• Language: English; English
• Published: 2016
• Subjects: TJ Mechanical engineering and machinery
• Online Access: https://eprints.ums.edu.my/id/eprint/18513/1/Towards%20Real%20Time%20visual%20Biometric.pdf; https://eprints.ums.edu.my/id/eprint/18513/7/Towards%20Real-Time%20Visual%20Biometric%20Authentication.pdf; https://eprints.ums.edu.my/id/eprint/18513/
• Institution: Universiti Malaysia Sabah

Telepresence Mobile Robots have prominent attributes in many fields as it provides virtual presence for human robot interaction. The deployment of this robot in healthcare sector has improved patient care and health. The vision system in a telepresence robot allows two way audiovisual communication between people at different location. In spite of such advancement, the manual way of controlling a robot to recognise and track people during an emergency is not favourable for a long duration. To circumvent this problem, biometric method using human face is proposed in this research which is implemented on Medical Telediagnosis Robot. This paper details the design of the face recognition and tracking system with four automated modules which are motion detection, face detection, face recognition and face tracking. The modules are developed with different algorithm and tested individually to ensure the stability of the system. Artificial Intelligence technique was applied at the face recognition stage while a two degree of freedom mechanism for actuator control was used at face tracking stage. A sequential mode operation is proposed to reduce the execution time in a real-time environment. To achieve this, only one module is operated at each time. A Graphical User Interface was developed to ease the users at the local and robot environment. The system is designed in LabVIEW platform. The biometric system proposed with hybrid algorithm at each module adapts for face images detected at different distances, poses and lighting condition. This system was tested in real-time and has an execution time of 55ms and 98% accuracy. The stand alone system designed for Medical Telediagnosis Robot can be will be very fruitful for various biometric system using facial technology.