Convolutional neural network based face recognition approach

Face recognition is the process of assignment of correct label to the face under consideration. Process of face recognition comprises of extraction of features from underlying face and feeding the extracted features to classifier to identify the corresponding individual. The Accuracy of classifier is greatly affected by the nature of features extracted. Conventional face recognition system relies on manually engineered, handcrafted features. Convolutional neural network is a deep learning model which automatically extract features from the raw data in the process of end to end classification. The automated feature extraction property of convolutional neural network not only saves the effort in manually extracting features but also solve the dilemma of set of features to be used for classification. In this work, we present a face recognition system based on convolutional neural network. We create our own dataset to test the efficiency of the proposed system. The accuracy of around 96% is achieved on the test dataset consisting of around 1900 images with 10 different classes.