The Classification System Of Cataract Fundus Image Based On Combinied Classifier

Cataract is the leading cause of blindness,early detection and treatment can reduce the pain of patients with cataract,prevent eye damage aggravated.At present,the early identification of cataract is still a difficult problem,we need to rely on an experienced ophthalmologist for diagnosis,and there is some error.Therefore,the development of an automated cataract diagnosis system has a broad application prospects in the early identification of cataract.On the basis of the analysis of the fundus images,a classification system of cataract fundus image classification system based on combination classifier is proposed,which can be divided into four categories:normal,mild,moderate and severe.In order to achieve automatic recognition of cataract fundus image classification,this paper has done the following research work:1.The method of image preprocessing is studied.In the fundus image preprocessing stage,color space transformation,histogram equalization,top-bottom hat transformation and filtering operation of the image are conducted.2.In the feature extraction stage,three independent feature sets,i.e.,wavelet feature,sketch feature and texture feature,are extracted from each fundus image.These three groups of features are 60 dimensions,23 dimensions,39 dimensions,respectively.3.In the MATLAB environment,for each feature set,two base learning models,i.e.,Support Vector Machine and Back Propagation Neural Network,are built.Then the ensemble methods,majority voting and stacking,are investigated to combine the multiple base learning models for the final fundus image classification.The best performance of the ensemble classifier is 83.9%and 84.5%in terms of correct classification rates for voting and stacking,respectively.The results demonstrate that the ensemble classifier outperforms the single learning model significantly.4.Taking the MATLAB GUI as development tool,to achieve the identification of cataract fundus images,processing and recognition system of cataract images was compiled.It is believed that this research can not only help doctors to diagnose,but also can improve the level of diagnosis and the people's living standard,and has good prospects for clinical application and social benefits.