A face attribute recognition method based on multi-instance and multi-label depth transfer learning

Abstract

The invention discloses a face attribute recognition method based on multi-instance multi-label depth migration learning, which comprises the following steps of preparing a face image data set, extracting a plurality of neural layer features of a depth convolution neural network migration model for each face image, and combining the neural layer features to form a multi-layer face feature; builinga network model for extracting multi-label relationship features, and training the parameters of the network model by using multi-layer face features as input and multi-face attribute tags as true values; designing a linear binary classifier for each face attribute, migrating the trained network model of multi-label relational features to the multi-face attribute classifier model as a feature extractor, and training each linear binary classifier by using face image data set. The method of the invention selects a transfer learning mode, rapidly and efficiently migrates a highly active transfermodel to a selected data set, builds a multi-tag relation characteristic model with simple training structure, and simultaneously trains a linear binary classifier of multiple human face attributes.

Description

technical field [0001] The invention belongs to a face attribute recognition method realized by applying deep learning in the field of computer vision. Background technique [0002] Face attribute recognition is an important subject, which can be applied to the research in the field of face recognition, such as face confirmation and face identification. The face and the surrounding area in the face image have many attributes, among which the attributes that are often considered are gender, hairstyle, expression, whether to wear glasses or a hat, and so on. It is more complex and difficult to study the face attribute recognition of face pictures in non-laboratory scenes, because many constraints are different, such as different face poses, occluders, different resolutions, and differences in brightness and color. But it is of great significance for the research of face attribute recognition in non-laboratory scenarios, because it is closer to the face pictures obtained in re...

AI Technical Summary

Problems solved by technology

Moreover, training is a very time-consuming process, and the parameters to be adjusted are not easy to determine in different situations, and the degree of execution in practical applications is not very high.

Images