Method and system for synthesizing human face image

Abstract

The invention discloses a training set-based method and a training set-based system for synthesizing a human face image. In the invention, a training set comprises a plurality of image pairs; and each of the image pairs is divided into many small image block pairs, each of the many small image block pairs comprises a small image block of a first type and a small image block of a second type. The training set-based method for synthesizing the human face image comprises the following steps of: receiving a test human face image of the first type; dividing the received human face image of the first type into a plurality of small image blocks; selecting a matched small block from the small image blocks of the first type in the training set for each of the plurality of small image blocks formed by dividing the test image; determining the small image blocks of the second type, which correspond to the plurality of small image blocks of the test image, in the image small blocks of the training set according to the selected matched small image blocks; and synthesizing the human face image of the second type by using the determined small image blocks of the second type.

Description

technical field [0001] The present invention relates to a method and a system for synthesizing human faces, in particular to a method and a system for synthesizing human face images based on a Markov network model. Background technique [0002] An important application of face recognition technology is to assist forensic identification. For example, automatically retrieving criminal suspects from a police department's photo database can help police quickly narrow down the suspects. However, in many cases, the police were unable to obtain photos of the suspects. The best substitute at this point is usually a portrait based on an eyewitness account. Therefore, it is very important to be able to automatically find photos in the photo database according to the portrait. However, there are great differences between portraits and photos, and the psychological process of generating portraits is relatively complicated, so recognition based on face portraits is much more difficult...

AI Technical Summary

Problems solved by technology

Its disadvantage is that local small blocks are transformed independently of each other on a single fixed scale, so some large face structures cannot be learned, especially the shape of the face cannot be learned

Images