Research And Application Of Virtual Fitting Model Recommendation Technology

In recent years,with the development of somatosensory interaction technology,the Kinect-based virtual fitting system provides a new option for customers to purchase clothing online.Kinect can accurately measure the human skeleton model and generate three-dimensional data of the human body to further recommend the model of the clothes for the user and use the visualization technology for virtual trial installation.However,the virtual fitting technology has been consistently in the past few years,mainly because its use effect has been restricted by technical bottlenecks,that is,it is impossible to accurately measure the size of the human body.Different people's body types are different,and the clothing models of different manufacturers are not standardized.The traditional data screening and matching means that the size of the clothes selected by the user is always unsatisfactory,and the user experience and the user repurchase rate cannot be improved.Based on the manual measurement method of the human body or the virtual fitting system scheme,the clothing store has accumulated a large amount of user body type information and purchased clothing information,which provides effective data support for user shopping information mining.In order to solve the problem of virtual fitting fit,to improve the customer purchase rate,using the statistical learning-based machine learning method,learn the fitting model recommendation model from the historical user body information and the purchased clothing information,and pass the Kinect sense.Measurement technology,real-time access to the real user's body size sequence,to recommend a suitable clothing model has become a promising choice.Through this strategy,consumers can obtain the size of clothing suitable for their size when purchasing clothing,increase the satisfaction and trust of consumers in online clothing,thereby enhancing the purchasing behavior of consumers.In this paper,combined with Kinect somatosensory technology,machine learning technology and Unity3 D augmented reality development technology,a set of artificial intelligence-based virtual fitting model recommended technical solutions and their applications are studied.The specific work content and innovations are as follows:(1)Based on Kinect somatosensory measurement technology,the fitting function of human body's three-dimensional attributes is constructed,and the main attributes of human body shape are obtained in real time,which reduces the tedious work of traditional manual three-dimensional human body measurement.(2)Combining with 3ds Max human body three-dimensional database,this paper used multi-classification technology,trained different machine learning algorithms,and compared the performance indicators of different machine learning model to select the best fitting model instead of the clumsy scheme of data selection and matching.(3)Using the Unity3 D augmented reality development tool,the virtual-reality fusion effect between the three-dimensional clothes model recommended by the model and the real-time two-dimensional human body image is realized,including clothes fitting, gesture changing and so on,which can show the fitness of virtual clothes for users in real time.