Face sketch-picture recognition method based on asymmetrical combined learning

Abstract

The invention discloses a face sketch-picture recognition method based on asymmetrical combined learning. The method comprises the steps of 1, obtaining a training sample set and a testing sample set;2, dividing a sample subset; 3, obtaining a training pseudo-sample set; 4, constructing an asymmetrical feature matrix; 5, calculating an asymmetrical combined learning matrix; 6, calculating the similarity value of a sketch sample and a picture sample; 7, obtaining a recognition result. By means of the face sketch-picture recognition method based on asymmetrical combined learning, a deep featurevector of a sample is extracted through a deep convolution network, the asymmetrical feature matrix is utilized to calculate the asymmetrical combined learning matrix, the asymmetrical combined learning matrix is utilized to calculate the similarity of sketches and pictures, and the pictures most similar to the sketches are found out to serve as the recognition result. By means of the face sketch-picture recognition method based on asymmetrical combined learning, the training pseudo-sample set is added in the training process, an asymmetrical combined learning method is utilized to increase intra-class information, and the pictures corresponding to the sketches can be accurately identified.

Description

technical field [0001] The invention belongs to the technical field of image processing, and further relates to a face portrait-photo recognition method based on asymmetric joint learning in the technical field of pattern recognition and computer vision. The present invention can retrieve and identify the corresponding human face photo through the drawn human face portrait, and then determine the identity corresponding to the human face. Background technique [0002] In the criminal investigation and pursuit, the public security department has a database of citizens' photos, combined with face recognition technology to determine the identity of the suspect, but in practice it is generally difficult to obtain the photos of the suspect, but the suspect can be obtained through the description of witnesses Sketch face portraits of people for subsequent face retrieval and recognition. Due to the great difference between portraits and ordinary face photos, it is difficult to obta...

AI Technical Summary

Problems solved by technology

The disadvantage of this method is that when encoding face portraits and face photos, the recognition accuracy is low due to insufficient encoding feature information.

The disadvantage of this method is that due to the difference in texture information between the face portrait block and the face photo block, there is information loss in the texture between the synthesized face pseudo-photo and the face photo, so the recognition accuracy is low.

Images