Application of Fuzzy Kernel C-Means in face recognition to identify look-alike faces

Abstract

Machine learning has been rapidly evolving and continuously developing. Many problems can be solved by machine learning. One of those is face recognition. There are many application of face recognition. One application that will be discussed in this research is face recognition to identify look-alike faces. Such application can be useful when dealing with huge data. Machine learning method that will be used is Fuzzy Kernel C-Means. This method is the modification of previous method Fuzzy C-Means. The kernel used is Radial Basis Function Kernel. Each image is characterised by its features. It is believed that reducing the number of features can also reduce the cost. Therefore, a feature selection method called Chi-Square was also used. Solving this problem in face recognition using Fuzzy Kernel C-Means resulted in quite high accuracies.