Tehnički vjesnik, Vol. 30 No. 6, 2023.
Izvorni znanstveni članak
https://doi.org/10.17559/TV-20230514000634
Unveiling Patterns and Colors in Architectural Paintings: An Analysis by K-Means++ Clustering and Color Ratio Analysis
Liang Zhang
; Tiangong University, Tianjin 300387, China
Yiqu Zhang
; Tiangong University, Tianjin 300387, China
Yumeng Wei
; Tiangong University, Tianjin 300387, China
Tao Zhang
; Beijing Institute of Archaeology (Beijing Academy of Cultural Heritage), Beijing 100032, China
Jian Zhang
; Tiangong University, Tianjin 300387, China
Jun Xu
; Tiangong University, Tianjin 300387, China
Sažetak
This study delves into the intricate world of patterns and colors found in architectural paintings within the illustrious Forbidden City. Through an in-depth analysis, we identified seven distinctive patterns, creating a pattern factor library that showcases five examples for each pattern category. To extract the color schemes of each architectural painting type, we employed the K-Means++ algorithm for secondary clustering. Utilizing both RGB and HSV color space models, we examined scatter diagrams and histograms for three specific architectural color paintings. The results revealed a balanced distribution of warm and cool colors across all three architectural painting types. The prevalent colors observed in the Forbidden City architectural paintings were red, yellow, cyan, and blue, exhibiting low levels of saturation and moderate to high levels of brightness, evoking a serene and luminous ambiance. Through color ratio analysis, we established traditional color names that corresponded to the extracted color values from each painting. Our findings suggest that the colors and patterns within the Forbidden City architectural paintings communicate a profound sense of tranquility and grandeur, aligning with the cultural and artistic values held during the Ming and Qing dynasties.
Ključne riječi
architectural paintings; color ratio analysis; k-means++ clustering algorithm; pattern extraction; the forbidden city
Hrčak ID:
309236
URI
Datum izdavanja:
25.10.2023.
Posjeta: 308 *