Costume collocation generation method based on generative adversarial network

Abstract

The invention provides a costume collocation generation method based on a generative adversarial network. The method comprises the steps of collecting costume item collocation of an online user to generate a training set and a test set; adopting the training set to train a generative adversarial network costume matching generation model; testing the trained generative adversarial network costume collocation generation model by adopting the test set; and verifying the test effect, and giving a structure and a training mode of the generative adversarial network costume matching generation model.According to the costume collocation generation method based on the generative adversarial network provided by the invention, for a relatively small data set, the generative adversarial network costume collocation generation model can solve the problem of insufficient recommendation caused by lack of enough alternative data; for a large data set, clothes most similar to the generated clothes pictures can be searched in the data set so as to solve the problem that the recommendation complexity is too high.

Description

technical field [0001] The invention relates to the field of clothing recommendation, in particular to a method for generating clothing matching based on a generative confrontation network. Background technique [0002] Fashion-related computer vision problems are attracting more and more attention these days as the fashion industry moves rapidly towards online business. Among them, the generation method of fashion clothing matching has become one of the hot research directions, and its task is to recommend another clothing that perfectly matches a given clothing item. The key to designing a generative approach to fashion collocations is to model the compatibility between fashion items. At present, the industry has carried out many explorations on this problem, such as distance metric learning, twin neural network and regression neural network and other methods. Despite their success in predicting compatibility, their application in real-life scenarios remains problematic....

AI Technical Summary

Problems solved by technology

When the given data set is small or limited, there may not be enough items as recommendation candidates, resulting in insufficient recommendation, for example, the difference in preferences of different users for different clothing collocations is not considered, the recommendation complexity is high, and the The matching relationship between various categories of clothing (such as tops, bottoms, shoes, etc.)

On the other hand, when the data set is large, generating recommendations will face efficiency problems due to the need to calculate the compatibility between each item. In addition, since most methods use deep neural networks, this will consume huge computing resources

Images