Face image processing method and device and server

Abstract

The embodiment of the invention discloses a face image processing method and device and a server. The method comprises the following steps: obtaining a to-be-processed human face image; inputting thehuman face image into a pre-trained convolution neural network model, and obtaining the classification data outputted when the convolution neural network model responds to the input of the human faceimage, wherein the convolution neural network model takes a loss function as a constraint condition, so as to enable the feature cosine value of each class of the classification data to approach to one; obtaining the classification data, and carrying out the content understanding of the human face image according to the classification data. The cosine value between a feature vector and a weight value of the loss function is enabled to approach to one, so as to achieve the convergence of inner-class distance. The convergence of inner-class distance enables the inner-class distance of the classification data to be increased, and the increase of the inner-class distance enables the classification data to be different more apparently. The increase of the robustness of data also can improve theaccuracy of content understanding.

Description

technical field [0001] Embodiments of the present invention relate to the field of image processing, in particular to a face image processing method, device and server. Background technique [0002] With the development of deep learning technology, the convolutional neural network has become a powerful tool for extracting face features. For the convolutional neural network with a fixed model, the core technology is how to design the loss function so that it can effectively supervise the convolutional neural network. Convolutional neural network training, so that the convolutional neural network has the ability to extract face features. In the prior art, the cross-entropy loss function of Softmax is mainly used to supervise the training of the convolutional neural network model. Among them, Softmax's cross-entropy loss function trains the ability of the network to extract features, uses the last layer of the network as the expression of the face, maps the face data to the co...

AI Technical Summary

Problems solved by technology

[0003] However, the inventors of the present invention found in the research that the feature extraction method of Softmax's cross-entropy loss function is a non-end-to-end method, which is simple and easy to implement, but due to the lack of restraint in the class data obtained by its training, As a result, the inter-class distance that determines the size of the cosine space distance is not obvious enough, and the discreteness of the classified data is poor, resulting in the inability to improve the accuracy of content understanding

Images