An Image Retrieval Method About Ancient Chinese Characters Within Adaptive Localed Area

Ancient Chinese characters normally have complex structures which are composed ofmany strokes. Different characters may show a similar shape which results in unsatisfactoryanswers for their image retrieving using the traditional methods. In this paper, an imageretrieval algorithm for images of ancient Chinese characters is proposed based on theadaptive local area of characters. First of all, according to the need of users, the regionsneeded to be matched are marked on the sample of ancient Chinese character image. In orderto avoid the possible problems of noise and losing some stroke, generated in the process ofselecting the local region, the region is corrected automatically. Then we match the image bycomparing the Euclidean distance of feature parameters between sample image and databaseimages. Finally, the ancient images which users need are found. The experimental resultsshow that the proposed method could improve the efficiency of correlative research as anauxiliary tool.