Age recognition model training method, face age recognition method and related device

Abstract

The embodiment of the invention relates to the technical field of face image recognition, and discloses an age recognition model training method, a face age recognition method and a related device, and an age recognition model obtained through training comprises an age classification model, an uncertainty model and an age regression model. In the first stage, an uncertainty model is adopted to convert the influence of noise interference on a first predicted age value output by an age classification model obtained by training into an uncertainty probability. And in the second stage, the training set is adopted to train the regression neural network, and the uncertainty probability corresponding to each face image is introduced into a loss function to obtain an age regression model. The introduction of the uncertainty probability is equivalent to telling an age regression model: the extracted features of each face image have certain probability uncertainty. Therefore, in the training process, the image characteristics with high uncertainty probability need to be learned emphatically, the influence caused by the noise can be reduced, and the generalization ability and accuracy of the model are improved.

Description

technical field [0001] The embodiments of the present application relate to the technical field of face image recognition, and in particular to a training method for an age recognition model, a face age recognition method and related devices. Background technique [0002] A face image contains a variety of face feature information, such as face shape, face skin condition, face expression, face features, face age, etc. Among them, face age is an important feature information. The field of face image recognition has been widely used. For example, some clients running on mobile devices have the function of face age recognition, wherein the client obtains face images and outputs the recognized face age based on the obtained face images to feed back to user. [0003] For these clients with face age recognition functions, the accuracy of age recognition, that is, the gap between the recognized age and the user's real age, is a more important content. At present, in the relevant...

AI Technical Summary

Problems solved by technology

[0004] The main technical problem to be solved by the embodiments of the present application is to provide a training method for an age recognition model, a face age recognition method and related devices, which can solve the technical problem of low accuracy in the existing face age recognition technology

Images