Structure adaptive CNN (Convolutional Neural Network)-based face recognition method

Abstract

The invention discloses a structure adaptive CNN (Convolutional Neural Network)-based face recognition method, and belongs to the field of identity recognition. According to the method, the advantages of a conventional CNN directly extracting characteristics from a two-dimensional image for recognition are maintained, and in addition, a network structure is adaptively constructed to overcome the shortcoming of excessive dependence of the conventional CNN on human experiences. According to the structure adaptive CNN-based face recognition method, the network is extended according to network requirements, so that the controllability and adjustability of the network structure are achieved, ineffective training is also avoided, difficulty in training for face recognition is lowered, and in addition, an optimal face recognition network structure is obtained; by adaptive network extension advantages, a newly added face sample can be relearned on the basis of maintaining early recognition results, retraining overhead is reduced, and incremental learning is implemented; the automatic intelligent face recognition method is low in training difficulty and high in accuracy under a big data condition.

Description

technical field [0001] The invention relates to a face recognition method and belongs to the field of identity recognition. Background technique [0002] As an important human biometric feature, face has important application value in intelligent identification. Because face recognition has the characteristics of directness, friendliness, convenience, covert operation, non-invasiveness, and strong interactivity, it has been widely concerned by academia and industry. Face recognition based on machine learning methods has attracted much attention in many face recognition methods. Among them, the shallow learning strategy represented by support vector machine and Boosting algorithm has greatly improved the performance of face recognition. [0003] In recent years, with the deepening of research on deep learning algorithms, Convolutional Neural Networks (CNN), which can be directly applied to two-dimensional image feature extraction and recognition, have gradually been applied ...

AI Technical Summary

Problems solved by technology

[0004] The network construction of traditional CNN relies heavily on empirical knowledge, and there is no theoretical knowledge based on parameter setting. Therefore, when determining the optimal structure of CNN, it is necessary to compare the performance of CNN with different parameter settings and different structures, and then adopt the best method. The high-performance CNN structure is used as the final network structure, which brings huge time consumption and hinders the further development of CNN under big data

In response to the above situation, many methods for improving the CNN structure have been proposed, most of which are based on the traditional CNN structure to build the network. In the end, the method still needs to determine the final network structure by comparing the performance of different network structures. The recognition training is difficult, the training time is uncontrollable, and automatic intelligent recognition cannot be realized.

Images