Research And Application Of Gray-scale Image Features Extracted Based On Fuzzy Clustering Of The Sample Area

Image feature extraction is an important component of computer processing image data. Due to the subjectivity of human vision, image is processed by the means of fuzzy. And the lack of training sample images needs to be analyzed by unsupervised methods. Fuzzy clustering analysis just meets the need of these two aspects. Thus this essay chooses fuzzy clustering method to analyze the gray level image features, and then realizes gray image feature extraction.This unsupervised analysis, also called cluster analysis, is the process that objects are distinguished and classified according to objects'similarity. While objects exist as the intermediary, so they are suitable for soft partition. Combining unsupervised feature and soft partition's feature of fuzzy cluster algorithm, this paper presents the image feature extraction method based on fuzzy cluster. This fuzzy cluster technique deal with the problem of similarity degree for finishing an optical image feature extraction processing using the method of similarity and statistics that used to calculate category object by establishing fuzzy relations. The image feature extraction based on fuzzy cluster presents significant advantages, which can complete the selection to the image region extraction or edge detection by adjusting system parameters.Firstly, this paper describes the theoretical analysis based on fuzzy clustering; secondly, the specific method of the image feature extraction is implemented using software programming ideas; finally, through entering some different system parameters, the image feature extraction system based on fuzzy clustering of the sample area is simulated. Comparing the results of image features with the result from the method of maximum and minimum, the proposed feature extraction method is verified correct and effective.