<![CDATA[Breast cancer is one of the deadliest diseases, especially for women. Early diagnosis of breast lesions and differentiation of malignant nodules from benign nodules and normal nodules are important for breast cancer prognosis. In diagnosing this disease, one radiological method, namely medical image analysis using ultrasonography, can be used to determine early diagnosis of breast cancer. Breast cancer ultrasound images have several characteristics, such as color, shape, size, and texture, which make segmentation difficult due to object accumulation. This study implements a Convolutional Neural Network classification algorithm and modified watershed segmentation to separate nodules or tumors in breast cancer. From the segmentation performance test with Watershed Transform, the average ZSI index was 40% for malignant images and 60% for benign images. The results of the VGG architecture for classification modeling showed 47% for watershed segmentation and 80% without watershed segmentation.]]>
Copyrights © 2025
