本文選題：聚類算法　切入點(diǎn)：初始聚類中心　出處：《絲綢》2017年12期

【摘要】：鳥類羽毛色彩進(jìn)化了上萬年,形成了一套天然美感的搭配法則,能夠?yàn)樵O(shè)計(jì)師提供設(shè)計(jì)靈感。為探尋其中的配色規(guī)律并加以利用,文章基于蜂鳥羽毛的色彩進(jìn)行了一系列的仿生色彩提取研究。首先針對K-MEANS聚類算法,對比分析了色相模式與灰度模式兩種不同的初始聚類中心的設(shè)置方法,并在此基礎(chǔ)上提出了一種經(jīng)過優(yōu)化的初始聚類中心的設(shè)置方法;然后使用優(yōu)化過后的K-MEANS聚類算法,對10種蜂鳥共200張羽毛色彩照片進(jìn)行羽毛色彩的主色系提取。通過三次聚類計(jì)算,探討分析蜂鳥羽毛色彩的配色規(guī)律,制得相應(yīng)的占比配色色卡,為設(shè)計(jì)師在以蜂鳥羽毛色彩為靈感來源的配色方案上提供了一定的理論依據(jù)。
[Abstract]:Bird feather color evolved for tens of thousands of years, forming a set of natural aesthetic collocation rules that can provide designers with design inspiration. In this paper, a series of bionic color extraction methods based on hue of hummingbird feathers are studied. Firstly, aiming at K-MEANS clustering algorithm, two different initial clustering center setting methods, hue pattern and gray model, are compared and analyzed. On this basis, an optimized initial clustering center setting method is proposed, and then the optimized K-MEANS clustering algorithm is used. In this paper, 200 feather color photographs of 10 species of hummingbird were extracted from the main color system of feather color. The color matching law of hummingbird feather color was analyzed by three cluster calculation, and the corresponding color matching card was obtained. It provides a theoretical basis for the design of color matching scheme based on hummingbird feather color.
【作者單位】： 浙江理工大學(xué)"紡織纖維材料與加工技術(shù)"國家地方聯(lián)合工程實(shí)驗(yàn)室;達(dá)利絲綢(浙江)有限公司;
【分類號】：TP391.41
，

本文編號：1664231