Penerapan Metode Jaringan Saraf Tiruan Propagasi Balik Untuk Mendiagnosis Tingkat Keganasan Kanker Payudara

Abstract

Medical imaging (Mammography) is very helpful in the breast cancer diagnostic. Mammography image analysis which is done manually requires expertise, experience, and subjective. A computer aided diagnostic system is helping radiologists to more objective, quicker, and more aqqurate detection of breast cancer. This research classifies benign and malignant breast cancer using backpropagation neural network (BpNN) with backpropagation method and gray level co-ocurence matrix (GLCM) feature extraction. Mammography image preprocessing methods Contrast Limited Adaptive Histogram Equalization (CLAHE) and Otsu’s N Segmentation. The object will be recognized cancer malignancy from spread of calcification in the breast. Feature extraction will be calculated at an angle of 00, 450, 900, dan 1350 and the number of hidden layer neurons tested were 4 and 8. The result is giving an average accuracy of 63.33% at angle of 90% and the number of hidden layer neurons is 4.