A REVIEW ON RESOURCE-EFFICIENT DEEP LEARNING APPROACHES FOR BREAST CANCER DETECTION USING ULTRASOUND IMAGING
Keywords:
Breast Cancer, Ultrasound Imaging, Deep Learning, Convolutional Neural Network, Transfer Learning, Computer-Aided Diagnosis.Abstract
Breast cancer remains one of the leading causes of death among women worldwide. Early and accurate detection plays a crucial role in reducing mortality rates. Traditional diagnostic tools such as mammography, while effective, are limited in dense breast tissues and inaccessible in low-resource regions. Deep learning, particularly Convolutional Neural Networks (CNNs), has emerged as a transformative technology in medical imaging. This review explores resource-efficient deep learning architectures and methodologies applied to ultrasound-based breast cancer detection. The paper discusses key CNN models, transfer learning strategies, and lightweight architectures such as DenseNet, VGG, and ResNet, emphasizing computational efficiency and diagnostic accuracy. A comparative analysis of recent studies from 2015–2024 is presented, highlighting advancements, limitations, and future research directions for cost-effective diagnostic AI systems. The review underscores the importance of developing resource-friendly CNN models suitable for real-world healthcare applications, especially in developing countries.
