"Benign and maligenant breast tumors classification based on the region growing and CNN segmentation" translation reading and understanding

Note: My English proficiency is limited, translation is inappropriate, please the original English, do not like to spray, the other, the translation of this article is limited to academic exchanges, does not involve any copyright issues, if there is improper infringement or any other other than academic communication problems, please leave a message I, I immediately delete, thank you!!

"Classification of benign and malignant breast tumors based on regional growth"

Summary

Benign tumors are considered to be one of the common causes of female mortality, and early detection of benign tumors can improve the patient's survival rate, so it is important to create a system that detects the suspicious tissues of the mammary gland. Two methods for automatic detection of benign and malignant tumors are presented in this paper, and the first method is to use the automatic region growth method to segment the graph, and the threshold value of regional growth method is obtained by Ann. In the second method, the Cellular neural Network (CNN) is used for image segmentation, and the parameters of CNN are obtained by genetic algorithm (GA). pixels, text and morphological features are extracted from the segmented mammary gland, and the GA algorithm chooses the appropriate features. In the next stage, the use of Ann to classify breast tumors, in order to evaluate the different classifiers (such as Random Forest, Bayesian, SVM,KNN) classification performance, in the MIAs and DDSM database tested, respectively, to obtain sensitivity, specificity, the overall progress of 96.87, 95.94,96.47.

"Benign and maligenant breast tumors classification based on the region growing and CNN segmentation" translation reading and understanding