Fuzzy Quaternion-based Pixel Wise Information Extraction for Face Recognition

In advancement of today's technological era, face recognition is an important task in various security systems, monitoring criminal and fraudulent activities. In this article, we present a new method for information extraction using two new concepts viz. quaternion and fuzzy logics. Quaternion helps in incorporating interrelationship between three color channels of a color face image. After obtaining quaternion vector of a color image we fuzzify it by using membership function (MF) i.e. π-MF, which find out in how much amount a pixel value of a color face image is associated with other face images of different classes. After that principle component analysis and k-nearest neighbor classification is used to obtain error rate in classification accuracy over Georgia Tech and Indian face (females) datasets. Investigational results clearly show the efficacy and superiority of our approach.