Intelligent Defect Detection Using Machine Learning to Improve Fabric Quality Control

The fabric defect identification process is a tedious and time-consuming process that involves multiple human work-forces to reduce the time consumption and to reduce human involvement we propose a novel approach applying Machine Learning techniques to identify the defects in fabric material. As per the global marketing data the textile business grows approximately 5 % to 6 % every year and reaches a maximum of one trillion dollars annually. With such a huge income-based industry the involvement of manual defect identification in fabrics is not cost efficient and defect identification efficiency is low. To overcome all these factors Machine Learning Algorithm are applied. We use a One-Class Support Vector Machine (OCSVM) to classify the defects based on features such as color intensity and design pattern. It automatically detects the defects in fabric by applying binarization and feature extraction techniques. These two techniques provide a rapid identification of defects in the fabric when compared to human identification with an accuracy more than 95%.