Face Recognition Using Convolutional Neural Network (CNN)

Abstract

Face recognition is one of the most important applications in video surveillance and computer vision. However, the conventional algorithms of face recognition are susceptible to multiple conditions such as lighting occlusion viewing angle or camera rotation. Therefor face recognition based on deep learning can greatly improve the recognition speed and compatible external interface. In this thesis we use convolutional neural networks (ConvNets) for face recognition the neural networks have the merits of end-to-end sparse connection and weight sharing. The purpose of this thesis is to identify the age, gender and ethnicity of different people based on location of the detected box of a face. Then we can obtain recognition result with different methods. Image processing is the wide area of studies and need highly integrated software and hardware tool. Currently a number of researchers are used with python integrated with OpenCV, Keras and Tensorflow and this proposed study is works in the same. The proposed research used convolutional neural network algorithms because it has speed when compared to other algorithms. The proposed algorithms used 19108 individual images with different features. Datasets are split in to training and other for Testing. The efficiency of the model in this research experiment is measured by deep learning approaches the confusion matrix is used to assess model performance. The TomNet model is tested in different epochs and learning rates to generate an efficient model. The result of the experiment is 98.88% training accuracy with 98.56% validation accuracy. The use of neural network for face recognition improves the speed of recognition. The contribution of this thesis is: can identify a human face including rotation and position. The confidence of human face recognition is mainly affected by the proportion of face on the screen.