Abstract
Breast cancer is the leading cause of death in women worldwide, the early detection can prevent possible deaths. The limited availability of breast ultrasound datasets prevents researchers from obtaining a good performance of the classification algorithms. Traditional augmentation approaches are firmly limited, especially in tasks where images follow strict standards, as in the case of medical datasets. Therefore besides the traditional augmentation, we use a new methodology for data augmentation using Generative Adversarial Network (GAN). We achieved higher accuracies by integrating traditional with GAN-based augmentation. I used two types of GAN models and compared them, Traditional GAN and DCGAN. This paper uses three datasets of breast ultrasound images acquired from three different ultrasound systems. The first dataset was taken from Baheya Hospital for Early Detection and Treatment of Women's Cancer, Cairo, Egypt. We name it BUSI, which stands for the Breast Ultrasound Images dataset, and it comprises 780 images: 133 normal, 437 benign, and 210 malignant. Dataset B is obtained from Spain from related work and has 163 images: 110 benign and 53 malignant. Last Dataset (online medical images dataset) consists of 197 images, with 136 being malignant and 61 benign.
