Approaches to the Task of Searching for Anomalies in Textile Texture Using Neural Networks

Tissue defect detection is a quality control process that aims to identify defects and determine their location. This process allows for the precise identification of defective areas and avoidance of them entering the finished product, which is of great importance for textile manufacturers. The ability to accurately pinpoint defect points to support a fabric quality control process is the primary goal of an automated patterned fabric defect detection and classification system. This should be achieved at the expense of good processing speed, less computational complexity, and less computation time. Thus, the designed systems require reliable and efficient algorithms for detecting textile defects. Although various types of fabric defects have been mentioned in the literature, only a few have been mentioned with patterned and colored patterned fabrics. Therefore, the purpose of this article is to present personal experience in the application of various approaches to detecting color defects in patterned fabrics using technical vision and machine learning technologies.