| 摘 要: 丝巾作为传统配饰享誉世界。对丝巾花型图案的风格进行分类,可以揭示丝巾花型之间的共性或差异,帮助企业和设计师有针对性地生产和设计产品。文章以丝巾为研究对象,提出了一种基于特征融合的丝巾花型艺术风格分类算法。首先,该算法使用风格迁移模型提取丝巾全局风格特征;其次,使用 SIFT(Scale-Invariant Feature Transform)算法提取丝巾局部特征;最后,使用特征融合算法对丝巾进行风格分类。实验结果显示,该算法在丝巾花型生成数据集上的分类总准确率高达91.90%,在实际丝巾产品数据集上的分类总准确率也达到89.06%。实验结果表明,该算法从多个角度对丝巾花型图案风格进行有效评估,并成功完成丝巾花型图案风格分类任务。 |
| 关键词: 花型图案生成;风格特征提取;风格分类;特征融合 |
|
中图分类号: TP39
文献标识码: A
|
|
| Artistic Style Classification of Silk Scarf Patterns Based on Feature Fusion Algorithm |
|
MAO Zhiyuan1, ZHANG Huaxiong
1, LIU Zhi1
|
(1.School of Computer Science and Technology, Zhejiang Sc-i Tech University, Hangzhou 310018, China;
2.Hangzhou Wensli Silk Culture Co., Ltd., Hangzhou 310021, China)
1104313695@qq.com; zhxhz@zstu.edu.cn; liuzhi@wensli.cn
|
| Abstract: Silk scarves are traditional accessories renowned worldwide. Classifying the styles of silk scarf patterns can reveal the commonalities or differences among them, helping businesses and designers to target their production and design efforts. This paper focuses on scarves and proposes an artistic style classification algorithm for silk scarf patterns based on feature fusion. First, the algorithm uses a style transfer model to extract the global style features of the silk scarf. Then, it employs the SIFT (Scale-Invariant Feature Transform) algorithm to extract the local features of the silk scarf. Finally, feature fusion algorithm is used to classify the styles of the silk scarves. Experimental results show that this algorithm achieves an overall classification accuracy of up to 91.90% on a silk scarf pattern generation dataset, and an overall classification accuracy of 89.06% on actual silk scarf product datasets. These results indicate that the algorithm effectively evaluates the styles of silk scarf patterns from multiple perspectives, and successfully performs the task of style classification for silk scarf patterns. |
| Keywords: pattern generation; style feature extraction; style classification; feature fusion |
