The Acquisition Of Human Body Size Based On Images

The human body size acquisition has a wide range of application scenarios,and has a very important role in the field of ergonomics,virtual fitting system,athletes' physical condition monitoring and medicine.With the development of clothing information technology,in order to achieve the purpose of personalized clothing customization,to obtain the user's personalized real human body size parameters,people began to seek a variety of non-contact human body size with both accuracy and efficiency.Some traditional human body size acquisition techniques require external tools for manual contact measurement,and non-contact measurement methods using scanners have the disadvantages of low efficiency and high cost.Based on the analysis of the research status and applications of the acquisition of human body size at home and abroad,this paper proposes an image-based human body size acquisition method.This method only needs to obtain the size of each part of the human body from the frontal photos,side photos,and background photos provided by the user.First,using the distance of the pupil of the human body as a reference,the initial size of each part of the human body is calculated by the proportional model,and then the personal height value manually input corrects the measurement error.The method first carries out pupil detection on the whole body frontal image to locate the pupil position,and calculates the pixel size of the human pupillary pitch.According to the feature that the adult pupil is a fixed size,the ratio of the pixel size to the true size of each part of the human body is calculated;The whole body frontal image provided by the user and the features of each part of the side image are located,and the pixel size of each part of the human body is calculated;the initial size of each part of the human body is obtained by the pixel size and ratio ? of each part of the human body;The height value is used as a regularization constraint term to obtain a regularized coefficient through data training to improve the accuracy of the measurement data.The method has the characteristics of high efficiency,real-time and without external tools;compared with some traditional non-contact measurement methods,the measurement cost is low and the measurement efficiency is improved.The research results can be applied to online clothing retail and clothing personalization.Customization and so on.