基于优化图像识别算法的自动剪纸工艺仪器设备研究

Research on Automatic Paper Cuttings Technology Instrument and Equipment Based on Optimized Image Recognition Algorithm

为了拓展传统剪纸工艺在自动化技术背景下的生存空间,本次研究将从图像分割和图像识别检索技术之中构建自动剪纸生成设备。该系统主要依靠Grabcut算法和改进的FREAK算法进行图像主体提取和特征检索。为了验证剪纸生成系统,本次研究通过Matlab平台进行仿真实验。在图像样本集中首先验证了不同算法的图像分割和图像检索性能,Grabcut算法在800张图片的分割中最低的错误率为37.2%,耗时14.1 s;同时,在剪纸设备仿真中,改进优化的SIFT-FREAK算法在不同纸张大小下精度分别为0.7431、0.7459、0.7466。结果表明,本次研究构建的自动剪纸生成系统在生产过程中精度性能高于其他算法。同时系统在人像景象主体分割和意象符号检索的剪纸生成之中均具备有效应用价值。

In order to expand the living space of traditional Paper Cuttings technology in the context of automation technology,this research will build automatic Paper Cuttings generation equipment from image segmentation and image recognition retrieval technology.The system mainly relies on Grabcut algorithm and improved FREAK algorithm for image subject extraction and feature retrieval.In order to verify the Paper Cuttings generation system,this research carried out simulation experiments through the Matlab platform.In the image sample set,the performance of different algorithms for image segmentation and image retrieval is first verified.The Grabcut algorithm has the lowest error rate of 37.2%in the segmentation of 800 images,and takes 14.1 seconds;At the same time,in the simulation of Paper Cuttings equipment,the precision of the improved and optimized SIFT-FREAK algorithm under different paper sizes is 0.7431,0.7459 and 0.7466 respectively.The results show that the precision performance of the automatic Paper Cuttings generation system built in this study is higher than other algorithms in the production process.At the same time,the system has effective application value in the subject segmentation of portrait scenes and the generation of Paper Cuttings based on image symbol retrieval.