Assistant human facial recognition based on convolution feature-map

Abstract

Face recognition is a very important research direction in the field of image processing. In daily life, there are various problems that require face recognition everywhere, but how to accurately identify faces from images is a very important and urgent problem. To this end, we propose an assistant face recognition method that can effectively improve the recognition accuracy. The face recognition method based on deep learning will convolve the input image into five different convolution kernels to obtain the convolution feature-map of each layer. At the end, the corresponding description vector of the input image is generated by using the two layers fully connected layer and the pooling layer. Therefore, in the process of face recognition, it is only necessary to calculate the similarity between the description vectors of the two frames of images. Because of its recognition accuracy and room for improvement, we use the features of the fifth-level convolution feature-map, combined with the image description vector finally generated by the deep learning method, can significantly improve the recognition accuracy in the face recognition process. Through a large number of experiments, the method can be used to assist the image recognition problem based on the deep learning method, which can effectively improve the recognition accuracy.