Design And Implementation Of Virtual Makeup Test System Based On Deep Learning

In recent years,with the rapid development of deep learning technology,its application in various fields has become more and more extensive,especially in the fields of face recognition and image processing.Therefore,more and more deep learning related technologies are used in the beauty industry,recommendation systems and other related industries,such as beauty and beauty software.However,for the virtual makeup application,most beauty software implementations are biased towards physical realization.The result of this implementation is that the makeup transition is unnatural,resulting in unrealistic makeup effects.Another point is that most of the existing beauty software functions are not perfect,and to some extent do not meet the expectations of users.In order to solve the above problems,this article designs and implements a virtual makeup system based on deep learning.This virtual learning system based on deep learning mainly designs and implements a role-based access module,a virtual makeup module based on common makeup types,a virtual makeup module based on a picture makeup template,and a personalized recommended makeup product module.For the sub-role access module,the module implements basic system functions such as login and registration,and provides basic judgment conditions for the recommended cosmetics module.For a virtual makeup test module based on common makeup types,this module mainly implements a segmentation model to recognize and segment different parts of the face.A makeup transfer model performs makeup transfer of different parts according to the segmentation results of the segmentation model.The main technologies used in the implementation are fully convolutional networks(Fully Convolutional Networks,FCN),VGG networks(Visual Geometry Group).For the virtual makeup test module based on the picture-like makeup template,this module mainly studies and applies the BeautyGAN technology to realize the instance-level virtual makeup test.For the recommended cosmetics module,this module mainly proposes and uses a dynamic parameter-adjusted recommendation algorithm based on the situation of the system user.The recommendation algorithm combines the advantages of the content-based recommendation algorithm and the user-based collaborative filtering algorithm and considers the system Features,dynamically adjust the weight according to the current user's record in the database to achieve more accurate and more satisfactory recommendations.Finally,the implementation and construction of the overall system is based on Python Web.The deep learning-based virtual makeup test system combines two different ways of virtual makeup test,presents a more complete makeup test function,and on this basis,realizes the function of personalized recommendation of makeup products,using different identities respectively The author uses different recommendation algorithms and methods to improve the accuracy of recommending makeup products and increase the user-friendliness and satisfaction of the system.Finally,the entire system has also passed the multi-thread high concurrency stress test.