Breast Architectural Distortion Detection Based On NSCT And Otsu-PCNN Algorithm

In recent years,the incidence of breast cancer has gradually become younger,which is a serious threat to women's health.Studies have shown that early detection can find lesions in time and reduce mortality effectively.In the field of breast cancer diagnosis and treatment,doctors diagnose lesions by checking the mammograms mainly,but with the increasing number of patients,the burden on doctors is increasing,misdiagnosis and missed diagnosis often occur,causing serious damage to patients.The computer aided detection(CAD)is of great help to improve the accuracy of breast cancer diagnosis,breast architectural distortion is a feature of the clinical manifestations of breast cancer,but the structure of breast architectural distortion is complex,the detection accuracy is low,and it seriously affects the diagnosis.Therefore,this paper proposes a new CAD scheme for breast architectural distortion.The doctor can mark the lesions of breast architectural distortion accurately in the mammograms by the scheme.In order to improve the accuracy of breast architectural distortion detection,according to the characteristics of the mammograms and breast architectural distortion,this paper studied a CAD method for breast architectural distortion.The main work and innovations of this paper are as follows:1.Aiming at the problems of the low contrast of the mammograms and the poor effect of the mammograms,this paper proposes a method that combines the top–bottom hat transformation with the gamma transformation for the pre-processing of the mammograms.First of all,this method excludes bright details smaller than the structural elements in the mammograms by the top hat transformation,thereby highlighting the foreground part of the mammograms,and then fills the trough information in the image by the bottom hat transformation,making the background part of the mammograms darker,and improving the contrast of the mammograms by the gamma transformation.The experimental results show that this method can effectively improve the contrast of the mammograms.2.In order to enhance the information of breast architectural distortion in the high frequency part and filter out the noise in the mammograms,this paper proposes to use an improved the non-subsampled contourlet transform(NSCT)to obtain the suspected lesion area of breast architectural distortion.The pre-processedmammograms are decomposed by the NSCT,for the high frequency sub-bands coefficients,the threshold is adaptively selected,and the coefficients large than the threshold are enhanced by a gain function,and set the coefficients small than the threshold to zero,through this process to enhance the information of breast architectural distortion and filter out the noise;for the low frequency sub-bands coefficients,the coefficients are processed by the linearly transformation to expand the contrast of the mammograms.Experimental results show that this method can effectively obtain the suspected lesion area of breast architectural distortion,making the detection of breast architectural distortion more accurate.3.In order to improve the accuracy of detecting breast architectural distortion and ensure the universality of the algorithm,this paper proposes to use Otsu to improve pulse coupled neural network(PCNN)to complete the segmentation of breast architectural distortion.The optimal threshold calculated by Otsu is set as the initial threshold of the PCNN model,which reduces the number of iterations of the PCNN.PCNN uses a simplified network structure,which reduces the impact of manual setting parameters and the impact of the parameters on the accuracy of the detection results.In order to ensure the universality of the algorithm,experimental tests are performed on the different density rating of the mammograms.The segmentation results are evaluated by sensitivity,specificity,accuracy,F1-score and the area under the ROC curve(AUC),the experimental results can fully demonstrate that the algorithm has high universality,and the experimental results also show that the CAD method of breast architectural distortion proposed in this paper has higher accuracy.