Face recognition method based on deep transformation learning in unconstrained scene

Abstract

The invention discloses a face recognition method based on deep transformation learning in an unconstrained scene. The method comprises the following steps: obtaining a face image and detecting face key points; carrying out transformation on the face image through face alignment, and in the alignment process, minimizing the distance between the detected key points and predefined key points; carrying out face attitude estimation and carrying out classification on the attitude estimation results; separating multiple sample face attitudes into different classes; carrying out attitude transformation, and converting non-front face features into front face features and calculating attitude transformation loss; and updating network parameters through a deep transformation learning method until meeting threshold requirements, and then, quitting. The method proposes feature transformation in a neural network and transform features of different attitudes into a shared linear feature space; by calculating attitude loss and learning attitude center and attitude transformation, simple class change is obtained; and the method can enhance feature transformation learning and improve robustness and differentiable deep function.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a face recognition method based on depth transformation learning in an unconstrained scene. Background technique [0002] Face recognition has been intensively studied for decades and applied in many aspects due to its non-invasive nature, and now many methods have been proposed to solve the face recognition problem. However, the change of human pose cannot be controlled, and it is often difficult to obtain frontal face images in unconstrained scenes, which is one of the main reasons for the low rate of face recognition or the inaccurate recognition of face images. [0003] In constrained scenarios, such as airport check-in and cash withdrawal from ATM counters (automated teller machines), it is easy to obtain frontal face images. In fact, there are many differences between the face images detected from the unconstrained scene and the constrained scene, and the detailed...

AI Technical Summary

Problems solved by technology

2. Nonlinear pose and shape changes

3. Mixed variation of different types of interference

Pal et al. proposed a dense keypoint framework to extract discriminative and kernelized features, and was able to handle non-unit transformations of faces, however, this unsupervised approach did not perform as expected.

Images