Study of the “background - product” color combination in the configuration of the GarmentScanner machine vision software and hardware complex

• Guseva M.A.

  Marina A. Guseva. The Kosygin State University of Russia. Moscow. Russia

• Yu. V. Rogozhina

  Yuliya V. Rogozhina. The Kosygin State University of Russia. Moscow. Russi

Technical vision technologies for recognizing objects from images are in demand in various sectors of the economy, healthcare, security systems, forensics, biometric identification, traffic analytics, logistics, etc. In the clothing industry, the introduction of machine vision began with noncontact anthropometry and digitization of patterns for sewing products. The results of the research in the field of automated recognition of fabric defects and seam defects in clothing contributed to the expansion of the scope of scientific developments. With the introduction of optical-electronic systems for quality control of semi-finished and finished products at sewing enterprises, a new advance in digital flaw detection will take place. At RSU named after. A.N. Kosygin a computer vision system that identifies defects in industrial batches of clothing is developed. Many factors influence the final result of the process of recognizing garment samples. One of these factors is the color combination in the “background – product” pair, which determines the accuracy of the program’s generation of the product’s icon outline. The article presents the results of an experiment on choosing a color scheme for the tabletop background of the GarmentScanner
computer vision software and hardware complex. It was developed as a virtual tool for measuring the dimensions of finished flat-shaped garment products.
Keywords: computer vision, scanning, recognition, color combination