Face recognition: A novel un-supervised convolutional neural network method

Image classification is an effortless task for humans but when it comes to learn by a machine it's fairly challenging. Convolutional neural network (CNN) as one of the most prevalent deep learning algorithm, has gain high reputation in Image features extraction. In this research article, we propose few new twists of unsupervised learning i.e. sparse filtering to seizure effective and distinguishable features of image. Features extracted by sparse filtering algorithm is convolved with the first CNN layer, and then these feature are further used in feed forward manner by the CNN to learn more good features for classification. The linear regression classifier is used to serve as the output layer of CNN for providing the probability of image class. We show that the performance of numeral visual identification and detection tasks improves by using these filters in multistage convolutional network architecture i.e. CNN. As far as we are concern, this is the first effort where an unsupervised convolutional neural network has been introduced to the facial classification problem. Experimental results on a public dataset determine the efficiency of the proposed model. Moreover, we conclude this paper by pinpointing some open research issues for face image classification in future contemplation.