Content-based Image Retrieval Of Chinese New Year Woodblock Painting

Chinese New Year Woodblock Painting is a kind of Chinese traditional folk art.It’s a symbol of Chinese New Year culture. With the development of the times,because of some reasons, Chinese New Year Woodblock Painting has faced extinction.How to research and protect Chinese Woodblock New Year Painting using computertechnology has become more and more popular.Content-based image retrieval uses image itself(or description of the image) asquery condition. It extracts features of images, compares the similarity of two imagesto measure the similarity of the two images. Then output the results according to thesimilarity. In this paper, we research and construct the system of Content-based imageretrieval. First, We extract features which applying to retrieval of Chinese New YearWoodblock Painting. The features include weighted and enhanced main colorhistogram, block color histogram, SIFT and global contrast based block salienthistogram. Then we compute the distance of the features as the measurement of thesimilarity. Finally, we assign weights to the four features and compute the value offinal similarity. We output the results according to the order of the similarity. This isthe most basic and important part of the system.As the same time, we construct index and relevance feedback mechanism toincrease efficiency, precision and recall. The two mechanism use SVM as classifier,which converts the two problems to classification problems. In addition, as furtherresearch, we discuss multi-image retrieval system and construct the system.Content-based image retrieval of Chinese New Year Woodblock Painting is useful tothe work of researching, protecting and inheriting the painting. It needs more researchand improvements to protect this Chinese traditional intangible cultural heritage.