Malignant Detection of Breast Nodules On BIRADS-Based Ultrasound Images Margin, Orientation, And Posterior

Breast cancer has the largest prevalence in world 2020, with 2,261,419 cases or 11.7%. It is also leading cause of death, accounting for 6.9% all deaths. Asia and Indonesia have greatest mortality rates. This an urgent issue that must be addressed. Ultrasonography (USG) advised assessing features breast nodules. nodules on ultrasound pictures are interpreted using Imaging, Reporting, Data System (BIRADS) category, which five features. Yet, probability a False Positive Result (FPR) imaging relatively high. Computer Aided Diagnosis (CAD) was created to reduce FPR. However, CAD research based many BIRADS traits currently margined. As result, three characteristics, namely margin, posterior, orientation aspects, this study aims proposed methode diagnosing nodule malignancy. The method consists 4 stages, namely, pre-processing, automatic segmentation, extraction, classification. Pre-processing adaptive median filter maximum window size 11 pixels, linear histogram normalizing, Reduction Anisotropic Diffusion (SRAD) were used construct method. neutrosophic watershed suggested segmentation. Based nodule's orientation, 10 proposed: width, gradient, slenderness, margin sharpness, shadow indicators, skewness, energy, entropy, dispersion, solidity. MLP classification approach. test 94 yielded accuracy 88.30%, sensitivity 82.35%, specificity 91.67%, Kappa 0.7449, AUC 0.865. it feasible conclude capable detecting malignancy images. To make more reliable future, RoI can developed.