Optimized Fabric Imperfection Detection via Image Analysis and Isolation Forest Technique

Fabric defect detection plays a crucial role in ensuring the quality of textile products. Traditional methods often rely on manual inspection, which is time-consuming and prone to human error. In this research, a novel method for fabric defect detection based on machine learning technique for anomaly detection the Isolation Forest algorithm is presented. The Isolation Forest algorithm builds a random forest of isolation trees to take advantage of the idea of isolating anomalies in a dataset. For the purpose of identifying fabric defects, every fabric sample is represented by a collection of features, including structural attributes, colour, and texture. By separating the majority of non-defective samples from the defective fabric samples, the Isolation Forest algorithm effectively finds the defective fabric samples. Based on the Standard Fabric Defect, our model also obtains a true positive rate value of 0.95 without generating any false alarms on our dataset.