Detection And Correction Of Back Defects In Men’s Suits Based On Image Processing

Men’s suits are a category of clothing that requires a high level of flatness and three-dimensional appearance,so it is important to judge and correct their appearance quality.This article aims to achieve the detection and correction of back defects in men’s suits using computer image processing technology and machine learning model training.The main research work completed is as follows:(1)First,using the image of a suit customization enterprise’s ills as the data source,a data set of five types of back ills was constructed,including back neck bulge,back too wide,shoulder blade bulge,back failure,and waist joint crack,with 10 pieces for each type,a total of 50 pieces.By analyzing four aspects that affect the fit of a suit,including the body shape structure,the pattern structure of a suit,the auxiliary materials for the suit surface,and the production process,the corresponding modification plans for each type of disadvantage are summarized.(2)The malady index parameters are extracted.First,use the EISeg image annotation tool of Feijiang to perform a matting operation on the image to obtain a preprocessing image that removes background information.Then,based on the MATLAB platform,the image is grayed out,and image processing methods such as image enhancement and threshold segmentation are used to draw the grayscale curve and binary image of the wrinkle area.The wrinkle width,wrinkle depth,and wrinkle density parameters characterizing the disease information are extracted based on the change degree of the peak and trough of the grayscale curve;Extracting wrinkle slope parameters based on the generated binary image;According to the different positions where the defects occur,the characteristic parameter of fold part number is added,and a total of five types of characteristic values are obtained.(3)Based on the BP neural network model,a genetic algorithm is proposed to optimize the super parameters of the traditional BP neural network model to solve the situation where BP neural network is prone to fall into local optimal solutions,and is applied to the classification of the types of back defects in suits.Divide 50 sets of feature parameters into training sets and test sets in a 7:3 ratio for classification and prediction,and output corresponding correction schemes.The simulation experiment results show that the model can accurately classify and identify the types of defects,with a classification accuracy of 97.1%,which is 3 percentage points higher than the training results of traditional BP neural network models.This article provides an effective solution for computer-aided identification and correction of back defects in customized suits by collecting actual cases of suit customization,analyzing the causes of defects,summarizing correction schemes,extracting feature parameters,and designing a BP neural network model based on genetic algorithm optimization.The experimental results show that the proposed method has high accuracy in classification accuracy and correction schemes,providing important support for the intelligent development of the customized suit industry,and providing a useful reference for the garment enterprise to detect defects in the ready-to-wear process after mass production of suits.