Research On Fabric Image Recognition Method Based On Fabric Attribute And Parameter Learning

With the development of the textile industry,the manual identification of cloth has been unable to meet the growing demand for production.More and more image recognition technologies are applied to cloth recognition.Image recognition is a technology that combines feature extraction and feature learning,which plays an important role in improving clothing industry competitiveness.But current clothing recognition algorithms usually have low recognition rate,because they only consider the vision attribute,which cannot fully describe the fabric and ignore the fabric properties of the fabric itself.Humans can recognize the world through tactile and visual perception,and machine learning can also benefit from this multi-modal perception ability to improve the recognition rate.In order to solve the problem of low recognition accuracy of common fabrics,this article focuses on the research of fabric image recognition from the aspects of fabric properties,tactile characteristics and parameter learning.At first,aiming at the problem that the current fabric image recognition relies on a single fabric attribute,the recognition rate is not high.This paper proposes a fabric attribute measurement method based on geometric measurements and tactile sensing to obtain two types of fabric attributes.The geometric measurement method is established to measure the input fabric image samples and the parametric modeling is obtained after quantitatively analyze the three key factors by testing the recovery,stretching and bending behaviors of different real cloth samples.The geometric measures of fabric properties can be obtained by parametric modeling.Besides,the fabric tactile sensing is measured through tactile sensor settings,and the adaptive features of tactile images are extracted by Convolutional Neural Network(CNN).This method comprehensively recognizes the fabric from the two visual and tactile perception dimensions of the fabric,avoiding a one-sided understanding of the fabric,and the two fabric attributes obtained can be effectively applied to subsequent fabric recognition experiments.Then,to solve the problem of low accuracy of fabric recognition,this paper proposes a fabric image recognition method based on fabric attributes and parameter learning.The fabric identification model is trained by matching the fabric geometric measures with the extracted features of tactile image and parameter learning through the convolutional neural network to learn the different parameters of fabric properties.Based on the AlexNet recognition model,a parameter learning module is added to complete the pairing of fabric geometric measures and tactile feature vectors,which completes the final identify model.The final experimental results show that the method in this paper effectively improves the accuracy and efficiency of fabric image recognition.Finally,based on the two methods above,connecting the current application scenarios and user needs of fabric image recognition,a fabric image recognition system framework based on fabric attribute and parameter learning is proposed,and the system prototype design is implemented.The system has a reasonable interface layout,basic functions,and accurate recognition results,which would better reflect the effectiveness and practicality of the method.