Chinese Herbal Medicine Feature Extraction And Identification System

Scientific identification of Chinese medicine is very significant, so this papertries to avoid using artificial means. Chinese Herbal Medicine inherent pictureinformation is used for scientific distinguish and identify, where it relates to acontent-based image feature extraction. Image level features mainly consist of threemain aspects, namely texture features, color features and shape features. This paperextracts three types of Chinese Herbal Medicine underlying characteristics and usesthe three types of weighted feature to classify and distinguish. When extracting thetexture feature, a patch fragment extraction method is be uesd, and this papercompares it with a filter extraction method; When extracting the color feature, afuzzy histogram method is be uesd and is compared with the traditional histogrammethod; Finally when extracting the shape feature, Hu invariant aquares are used torepresent shape feature.This paper achieves a system to analysis and recognition Chinese Herbalfeature, in the system, the three types of Chinese Herbal Medicine underlyingcharacteristics are be extracted, and this paper finishs feature integration for thethree types of Chinese Herbal Medicine underlying characteristics. This paper alsoimplements data storage, making use of the ADO database storage.In this paper, Chinese Herbal recognition has made breakthrough progress forfive kinds of datas in feature extraction and matching identification, and it hasaccessed high matching rate of98%.