Enhancing Ovarian Cancer Detection: A Deep Learning Approach with MobileNetV3 and ResNet50

Ovarian cancer, commonly known as the “silent killer,” presents notable obstacles in terms of timely detection and management. This work aims to explore the capabilities of deep learning models, namely MobileNetV3 and ResNet50, in improving the detection accuracy of ovarian cancer. Using an extensive collection of ovarian tissue photos, we performed a comparative examination of two advanced convolutional neural networks (CNNs) to assess their efficacy in discriminating between cancerous and non-cancerous tissue samples.The findings of our study indicate that MobileNetV3 and ResNet50 exhibit considerable potential in the identification of ovarian cancer. The MobileNetV3 model had a notable accuracy rate of 96.3%, highlighting its effectiveness in the early detection of cancer. The ResNet50 model exhibited notable performance, with an accuracy rate of 92.08%. The aforementioned results highlight the potential of deep learning models in raising the precision and effectiveness of ovarian cancer detection, a crucial measure in improving patient outcomes. The comparative analysis presented in this paper is a significant resource for healthcare practitioners and researchers, as it provides insights into the strengths and limits of these models. This analysis also lays the groundwork for future developments in the field of gynecological malignancy diagnosis.