AI-based structural and motif rule extraction for Weifang kite analysis and digital preservation
Аннотация
The famous Weifang kites can be taken as the symbol of the most perfect specimen of the China cultural craftsmanship with their detailed motifs, well-balanced geometric patterns, and colors symbolizing some certain cultural significance. However, the systematic computation analysis of such design principles has been little explored and particularly in quantifying the interaction of structural geometry, motif formation and color relationships to create coherent visual rules. This limitation limits mass digital catalog, rule-card generation of cultural preservation. In order to respond to this requirement, this paper presents an AI-based analytical model that will be used to combine image preperocessing, structural feature extraction, Convolutional Neural Network (CNN) based motif recognition, color pattern clustering, and rule generation. The processed images are converted into structural descriptors, which include wing angles, symmetry ratios, contour vertices and aspect ratios. A CNN separates the main motifs as dragon, bird, butterfly, fish, tiger, and human basing on hierarchical texture and contour data. These rules in structure, color, and motif are then incorporated into an automated engine of generating digital rule cards. structural feature detection is 97-percent, symmetry estimation is stable with a variation of 2.1 with confidence levels of over 90 percent of the significant rule categories.