One of the hot and popular topics in the today’s modern world is face recognition. It is especially useful in security issues for identifying people and finding their background. In general, in the process of diagnosis, images or features are usually converted to vectors. These methods usually leads to distortion in the correlation information of the elements in the removal of an image matrix. In this article, while introducing a new recurrent convolutional neural network (RCNN), it is used to face detection and recognition in color images. The use of the radial basis function neural network (RBFN) as well as the application of feedback in it, has led to the creation of a very strong convolution neural network for face recognition. The recurrent convolutional neural network first receives the image database as a 3D matrix and, after training, selects the closest face with acceptable accuracy. A comparison between recurrent convolutional neural network and traditional convolutional neural network has been done in the experimental analysis. There is also a comparison in the recognition rate between our proposed method and a number of recent articles. The experimental results show the efficiency of the proposed recurrent convolutional neural network.
