Developing Prototype for Prosopagnosia Using PCA

TLDR: The prototype system is trained using 30 candidates’ images including 10 images per candidate under distinct angles and was able to achieve 98% face detection, and can be trained at any point of time.

Abstract: Prosopagnosia primarily known as face blindness is a brain disorder in which a person is not able to recognize faces. In the device a portable camera is used to capture the real time movement of a person. In the main system face is detected using Voila –Jones Algorithm & hue channel detection. Using Principal Component Analysis (PCA) face is recognized. It is compared with the trained database and displays the basic information about the person if a similar trained data is found otherwise, it is declared as unknown. For new database, system can be trained at any point of time. The prototype system is trained using 30 candidates’ images including 10 images per candidate under distinct angles. The system was able to achieve 98% face detection. It provides the patients an aid to help them recognize people in front of them in seconds.