DIGITAL RECONSTRUCTION OF TEXTILE ORNAMENTS BASED ON NEURAL NETWORK IMAGE ANALYSIS

Abstract

The article explores the potential of applying neural network–based computer vision methods to the analysis and digital reconstruction of historical textile ornaments. The relevance of the study is обусловлена the fragmented preservation of textile artifacts, the loss of integrity of decorative elements, and the complexity of reconstructing ornamental structures using traditional approaches. The use of digital technologies increases the objectivity of visual fabric analysis and enables the automation of image processing procedures.

The research substantiates an approach to the application of computer vision algorithms for the recognition and reproduction of ornamental motifs based on digital images of historical textiles, with further use of the results in digital clothing design. The paper summarizes contemporary applications of deep learning in tasks related to texture structure analysis, classification of visual elements, and processing of complex decorative compositions.

The principles of constructing a multi-stage digital image processing workflow are outlined, ensuring a consistent analysis of textile ornaments and forming the basis for their further reconstruction. The role of neural network models in detecting structural ornament elements, identifying symmetry patterns, and forming repeating compositional units is emphasized. The source base of the study includes digital archives of museum collections, historical illustrative materials, and digitized fabric fragments.

The practical significance of the research lies in expanding the possibilities of digital preservation of cultural heritage and creating conditions for integrating historical ornaments into digital clothing design environments, particularly CLO 3D. The proposed approach establishes a foundation for the further development of digital textile design and the implementation of artificial intelligence tools in design practice.