|
|
|
本文已被:浏览 1788次 下载 1215次 |
 码上扫一扫! |
|
瑶族纹样符号的智能分类方法 |
孔谦,史卓,冯业,杨铭,张梦雪,曾树珍,李荣斌,玉珂,申婧
|
桂林电子科技大学,桂林 5410004
|
|
摘要:目的 针对人工分类方法繁琐且耗时的问题,依托人工智能技术对纹样符号进行智能分类,能够在减少耗时问题的同时,有效提高其分类的精确度,从而促进对少数民族民俗文化的传播。方法 首先从符号学及传播学角度分析智能分类的价值,然后通过文献研究、田野调查等方法收集瑶族服饰、织锦上的纹样符号并制作数据集,然后对其进行初步分类和预处理,最后采用CNN算法和Faster R-CNN算法分别对数据集进行训练和测试。结论 经过训练和测试表明,两种算法均能实现瑶族纹样的智能分类,并且识别的平均准确率较高。其中基于Faster R-CNN算法的智能分类方法可以对图片中的多种纹样进行定位和细化分类,更适合应用到实际场景中,且能够与新媒体传播方式相结合,借助民俗符号阐释和传播瑶族民俗文化,这对少数民族非物质文化遗产的保护和传承有着重要的价值和意义。 |
关键词: 纹样符号 瑶族服饰 瑶锦 图像识别 目标检测 智能分类 |
DOI:10.19554/j.cnki.1001-3563.2021.10.034 |
分类号:TB472 |
基金项目:广西高等学校千名中青年骨干教师培育计划人文社会科学立项课题《智能媒体语境下的广西少数民族符号非遗保护与传播研究》(2020QGRW017) |
|
The Intelligent Classification Method of Yao Patterns |
KONG Qian, SHI Zhuo, FENG Ye, YANG Ming, ZHANG Meng-xue,ZENG Shu-zhen, LI Rong-bin, YU Ke, SHEN Jing
|
(Guilin University of Electronic Technology, Guilin 5410004, China)
|
Abstract:In view of the complexity and time-consuming problem of artificial classification, the work aims to use the artificial intelligence technology to classify the pattern intelligently, soasto effectively reduce the time-consuming and improve the accuracy of classification, which can also further promote the spread of the connotation of pattern symbols and ethnic minority folk culture. First of all, this paper analyzes the value of intelligent classification from the perspective of semiotics and communication, and then, collects the pattern symbols on Yao costume and Yao brocadethrough literature research and field investigation to make a dataset. Finally, the data set is trained and tested by CNN algorithm and Faster R-CNN algorithm respectively. The training and testing show that, both algorithms can realize the intelligent classification of Yao patterns, and the average accuracy of recognition is high. Among them, the intelligent classification method based on Faster R-CNN algorithm can locate and refine the classification of multiple patterns in the picture, which is more suitable for the actual scene. It can be combined with new media communication to further explain and spread the Yao folk culture with the help of folklore symbols, which has great value and significance for the protection and inheritance of intangible cultural heritage of ethnic minorities. |
Key words: pattern symbol Yao costume Yao brocade image recognition object detection intelligent classification |
|
|
|
|