| Minority costumes are the characteristic minority culture formed in the long development of minority culture,which is a very valuable resource of our country.However,with the development of social modernization and economic globalization in recent years,minority cultural resources are on the verge of extinction,and even some have disappeared,the protection of the situation is grim.It is of great academic significance and application value to analyze the underlying visual features of ethnic costumes from the perspective of computer vision.In this paper,the main contents of the research on the digital image processing of minority costumes are as follows:(1)Image collection and image database construction of minority costumes.The costumes of this experiment are from the self-built database of ethnic minority costumes,and the source of clothing is our research team rushed to the Yunnan minority ethnic villages or museum real shooting.At present,the database included WA,Hani and Yi and so on image many of costumes,and database of the type and quantity of clothing are still building.(2)Improvement of fuzzy C mean clustering segmentation algorithm based on minority costume image.In this paper,Based on block truncation theory,an improved algorithm for image segmentation based on spatial neighborhood of Fuzzy C-mean algorithm is proposed.Firstly,the block truncation coding theory is used to cut the image R,G,B color space into six components,Then,the six dimensional feature vector is used to express the characteristics of the national costume image.Finally,it is used as the data input of the Fuzzy C-mean segmentation algorithm based on the spatial neighborhood.Comprehensive quantitative metrics and visual effect,experimental results show that the algorithm not only to image noise has good robustness,and on national dress image segmentation effect is better,especially on national costume has representative elements of the characteristics of the regional segmentation effect is remarkable.(3)Establishment of Image Segmentation System for Minority Costumes.In this paper,we use the technology of Java and Matlab mixed programming to construct the segmentation system of minority costume image,and solves the key technical problem of carrying on the parameter mutual transmission with the image as the parameter in the mixing.Thus,the system has the reliable stability of the Java language,the strong GUI design and the characteristics of the object-oriented,and has the powerful function of Matlab language graphics and image processing,as well as the function of numerical matrix calculation.In addition,the system not only integrates the algorithms proposed in Chapter 3,but also integrates a variety of classic and effective image segmentation algorithm,which provides the technical reference for the national costume element database construction and national costume related researchers. |