Face image generation method and device

Abstract

The invention belongs to the technical field of face recognition, and in particular relates to a method and device for generating a face image. It aims to solve the technical problem of how to augment the data of face image samples containing specific attribute information. The invention provides a method for generating a human face image, comprising inputting a human face sample image and a preset human face multi-attribute condition vector into an adversarial generation network; Sampling, the sampling results are input into the generation network of the confrontational generation network to obtain a synthetic image; respectively, the synthetic image and the face sample image are input into the discriminative network of the confrontational generation network, and when the discriminative network cannot distinguish the synthetic image from the input image, the synthetic image is used as Face image output that meets the face multi-attribute conditions. The present invention can expand the original generative confrontation network under the condition that the training set samples are difficult to obtain, and generate images with specified attribute conditions through the guidance of multi-attribute conditions.

Description

technical field [0001] The invention belongs to the technical field of face recognition, and in particular relates to a method and device for generating a face image. Background technique [0002] With the development of neural network technology, neural network technology is being applied in more and more scenarios, and face recognition, as the most active branch of biometric recognition, is receiving more and more attention. Among them, image and video retrieval and face attribute analysis are important applications of face recognition. Taking age estimation in face recognition as an example, in the traditional age estimation technology, due to the lack of face image sample data, it can only be based on Age distribution mining with limited data, or after obtaining image information of different regions of the face, age estimation is performed by combining the acquired image information. At present, face image samples can be expanded by using data augmentation methods, but...

AI Technical Summary

Problems solved by technology

At present, face image samples can be expanded by using data augmentation methods, but conventional data augmentation methods cannot expand face image samples containing attribute information such as specific age and gender, thus affecting the accuracy of face recognition. Thus restricting the application and promotion of face recognition

Images