Image data processing method and device, computer equipment and storage medium

Abstract

The invention discloses an image data processing method and device, computer equipment and a storage medium. The method comprises the steps of inputting N label images and M label-free images into animage recognition model; determining a first identity prediction result of the object contained in each label image in an image recognition model, and obtaining a similar image of each label image from the image set, wherein the image set comprises N label images and M label-free images, the similar image of each label image does not carry the object identity label carried by the label image; determining a similar difference loss function according to each label image and the similar image to which the label image belongs, and determining a prediction loss function according to the first identity prediction result of each label image and a carried object identity label; and correcting model parameters of the image recognition model based on the similarity difference loss function and the prediction loss function to obtain a target image recognition model. According to the invention, the accuracy of the target image recognition model for object identity recognition can be improved.

Description

technical field [0001] The present application relates to the technical field of image data processing, and in particular to an image data processing method, device, computer equipment and storage medium. Background technique [0002] With the continuous development of computer networks, AI (artificial intelligence) has been applied in all aspects of life, such as training models through machine learning in artificial intelligence, and then using the trained models for image recognition. [0003] In the prior art, it is usually possible to pre-set the expected global classification boundary according to the category label of the sample data, so that the model can gradually learn the feature space of the sample data according to the boundary. However, training the model by presetting the expected global classification boundary usually makes the trained model biased towards certain category attributes. For example, in the face recognition scene, the trained model will be bias...

AI Technical Summary

Problems solved by technology

For example, in the face recognition scene, the trained model will be biased towards the skin color attribute, hairstyle attribute or jewelry attribute in a certain type of face, and when the model is biased towards these types of face attributes, it will make the model in When performing face recognition, it is more inclined to recognize such face attributes, resulting in low recognition accuracy for global face attributes

Images