Research On Recognition Methods Of Breast Cancer Image

Early detection of breast cancer is the most effective way to save the lives of patients.At present,mammograms are the most commonly used for diagnosing breast cancers,but the specificity of the features in the initial mammograms is not strong.Affected by subjective factors of doctors,it is prone to misdiagnosis and missed diagnosis.At present,with the rapid development of computers,the advanced computer technology is used to help doctors diagnose breast cancers.This paper proposes new recognition algorithms based on the pattern recognition method of mammograms,and realizes the effective recognition of breast cancers.The details and innovations are as follows:(1)The related techniques of mammograms processing are introduced.The contrast enhancement algorithm is used to enhance the effect of the images and improve the images quality.Through the fuzzy c-means algorithm,the potential tumor area is effectively segmented.The characteristics of breast mass and breast density were extracted as effective features.In the area of breast mass,48 texture features and 5 shape features were extracted using the gray level co-occurrence matrix texture analysis method;2 gray scale features were extracted in the breast density region,and take the extracted features as the experimental data set of the following algorithm.(2)A parameter optimization algorithm based on support vector machine is proposed.To optimize the penalty coefficient and kernel function parameters in the SVM classification algorithm for breast cancer recognition,an improved fruit-fly optimization algorithm is proposed as an optimization algorithm for SVM parameters to improve classification performance.By establishing experiments,the classification results show that the algorithm can improve the recognition accuracy.(3)A joint enhancement coupled feature representation method for mammograms recognition is proposed,a coupled feature representation by utilizing intra-coupled and inter-coupled interaction correlationship is constructed,and a breast cancer recognition algorithm based on the joint enhanced coupling feature representation was established.