An Innovative Machine Learning (ML) Approach in Fabric Defect Detection and Quality Assurance

Aug 3, 2026

Fabric defect detection is a complex requirement for textile quality management, directly affecting product functionality, aesthetics, comfort, and durability. Traditional manual inspection suffers from fatigue, limited focus, and long processing time, increasing cost and reducing consistency. This study proposes an automated garment inspection system based on supervised machine learning using SVM, trained and evaluated with a 500-image dataset. Results report 72% precision/74% recall for holes and 85% precision/83% recall for stains, cutting inspection costs by 65% and improving productivity.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Introduction

Textile quality control and the role of defect detection

Limitations of manual inspection

Motivation for ML-based garment inspection

Proposed approach (overview)

Dataset and training

Supervised model (SVM) for fault classification

Performance for hole and stain defects

Scalability and implementation considerations

An Innovative Machine Learning (ML) Approach in Fabric Defect Detection and Quality Assurance Nida Khalil1, Khalid Mahboob2, Mustafa Ahmed Khan2, Qurat-ul-Ain Nayar2, Aimen Qasim1, Shayan Faiz2 1Sir Syed University of Engineering and Technology, Karachi, Pakistan 2Institute of Business Management, Karachi, Pakistan *[Correspondence](Correspondence: nkhalil@ssuet.edu.pk)[: ](Correspondence: nkhalil@ssuet.edu.pk)[nkhalil@ssuet.edu.pk](Correspondence: nkhalil@ssuet.edu.pk) Citation | Khalil. N, Mahboob. K, Khan. M. A, Nayar. Q. U. A, Faiz. S,“An Innovative Machine Learning (ML) Approach in Fabric Defect Detection and Quality Assurance”, IJIST, Vol. 07 Issue. 04 pp 2247-2262, October 2025 Received| August 21, 2025 Revised| September 28, 2025 Accepted| September 30, 2025 Published| October 02, 2025. The garment and textile industries nation's economic development. textile and technology industries are essential sectors that significantly contribute to a Fabric defect detection is a complex problem in the since the efficacy and efficiency of automatic defect detection determine the quality and cost of any textile product. In the past, the textile industry used manual labor to find flaws in the fabric production process. The primary disadvantages of the manual fabric flaw identification technique are human weariness, lack of focus, and time consumption. This article introduces an innovative automated system for detecting garment defects powered by machine learning to revolutionize the traditional system and replace the manual inspection system. This innovative advanced system is trained and assessed using the 500-image dataset from the Artistic Milliners Company in Pakistan. The machine learning algorithm and image processing techniques form the foundation of AI technology, offering the best flaw detection accuracy. This work presents an automated fabric defect detection system driven by a supervised machine learning algorithm, i.e., SVM, that can accurately and precisely detect "hole" and "stain" faults. The system achieves a 72% precision and 74% recall for holes and an 85% precision and 83% recall for stains by utilizing a machine learning algorithm, i.e., SVM. The proposed method throws up vital issues like scalability and fabric sort flexibility. Compared to traditional manual processes, this new method lowers inspection costs by 65%, increasing productivity and setting a standard for automated and sustainable textile quality monitoring. Keywords: Fabric Defects Detection, Hole Defect, Stain Defect, Machine Learning, Image Introduction: Among the various industrial sectors worldwide, the textile industry holds a central position in the global economy [1] . As international competition intensifies and consumers demand higher standards, maintaining rigorous quality control has become increasingly vital for the sector [2] . The quality of fabric plays a pivotal role in determining the end product’s functionality, aesthetic appeal, and overall profitability. Fabric defects, such as color inconsistencies, holes, thin spots, stains, and broken threads, not only reduce the visual appeal and functionality of the final product but also compromise its comfort and durability [3] . In order to guarantee both product quality and production efficiency, it is therefore becoming more and more important to identify and fix fabric flaws as soon as possible [4][5] . The ultimate objective of this endeavor is to create a machine learning–based garment inspection system that will boost the textile industry's overall production, accuracy, and efficiency. Detecting fabric defects has become a critical component of textile quality management, garnering a lot of interest and study [6] . Over the past few years, considerable efforts have been made to modernize and simplify garment inspection procedures. In traditional methods, manual labor has often been used, leading to wrong judgments, potential human errors, and fewer inspec

FAQs

Why is automated fabric defect detection important in textile quality assurance?

Fabric defects reduce visual appeal, functionality, comfort, and durability, while also impacting production efficiency and costs. Automated detection improves consistency and supports faster quality monitoring than manual inspection.

What machine learning method is used for defect detection and classification?

A supervised machine learning approach using SVM is employed to detect and classify fabric defects, specifically holes and stains.

What performance metrics does the proposed system achieve?

The system achieves 72% precision and 74% recall for holes, and 85% precision and 83% recall for stains, based on the evaluated dataset.

Aug 3, 2026

Upload by Genevieve