A vision-based approach to mapping flexible objects for garment development

Abstract

For variable and flexible objects, there is no appropriate intelligent method to quantitatively characterize the three-dimensional (3D) form, especially for garment development. To address the problem, we proposed a novel approach to mapping 3D flexible objects with the coded graphic as a medium. Two-dimensional mapping patterns were used to characterize the 3D form and extract metric information. The proposed graphic code is small in size and it is easy to demonstrate position. With different fabrication techniques, various coding materials are available. With only a monocular image, the method shows high accuracy and low cost without the need for camera calibration in advance. Specifically, the processes of the method, including the algorithm of feature extraction, decoding, mapping position calculation, and pattern generation, are discussed. Two tests were implemented, and the results showed that the method was accurate and simplified the process of made-to-measure garment development. The proposed method has great application potential in the manufacturing of labor-intensive and experience-dependent flexible industries, such as apparel, home decoration, shoes, and other related areas. It also sets the stage for further artificial intelligence research of flexible objects.