The COVID-19 outbreak has made it unsafe to deploy contact-based biometric verification for attendance marking systems. To circumvent this problem, Research Centre Imarat (RCI), DRDO, has successfully developed AI-based Attendance Application (AINA) which allows non-contact based personnel verification using facial features of the person, captured with the help of a low-cost camera.
AINA can be deployed with minimal upgrade to the legacy attendance infrastructure with RFID readers. It can be installed on a desktop computer with a GPU-based display adapter, said an official spokesman.
Existing CCTV cameras can be utilised for capturing facial images. Facial features of several thousands of employees can be stored in the computer since for each employee the facial features are encoded in a small (less than 25 KB) file.
The system is fully scalable since the time for identification and verification for each person remains constant even as the number of registered personnel increases. It is fully secure as it works as a standalone system and does not require any connection to the internet. Since only the facial features are saved in an encoded form, the actual face images need not be saved, thereby ensuring privacy and security.
Confined database
Besides, the server storing the facial feature database is confined to the organisation premises. AINA has a light-weight installation process and comes with a very intuitive and user-friendly GUI with audio prompts which can be optionally disabled. It has been currently installed in ESIC Hospital, Hyderabad,