Automated Fabric Defect Inspection: A Survey of Classifiers

TL;DR

This survey reviews various classifiers used in automated fabric defect inspection systems, highlighting their techniques, performance, and potential for improving accuracy and efficiency in textile quality control.

Contribution

It provides a comprehensive overview and comparison of classifiers for automated fabric defect detection, aiding researchers in selecting suitable methods.

Findings

Different classifiers have varying performance metrics.

Automated systems can improve accuracy and reduce inspection time.

Survey helps identify promising techniques for future research.

Abstract

Quality control at each stage of production in textile industry has become a key factor to retaining the existence in the highly competitive global market. Problems of manual fabric defect inspection are lack of accuracy and high time consumption, where early and accurate fabric defect detection is a significant phase of quality control. Computer vision based, i.e. automated fabric defect inspection systems are thought by many researchers of different countries to be very useful to resolve these problems. There are two major challenges to be resolved to attain a successful automated fabric defect inspection system. They are defect detection and defect classification. In this work, we discuss different techniques used for automated fabric defect classification, then show a survey of classifiers used in automated fabric defect inspection systems, and finally, compare these classifiers by using performance metrics. This work is expected to be very useful for the researchers in the area of automated fabric defect inspection to understand and evaluate the many potential options in this field.