Tripath Ensemble Architecture of Identifying Benign and Malignant Ovarian Cancer

Women's health is seriously threatened by ovarian cancer. It originates from the ovaries, which are crucial for the production of hormones like progesterone and estrogen as well as eggs. Usually, aberrant ovarian cell division results in the development of malignant tumors, which in turn causes ovarian cancer. It might be challenging to detect ovarian cancer in its early stages. However, addressing this sickness requires knowledge of the detectable techniques available. Using medical imaging, deep learning is very good at identifying and differentiating between a wide range of disorders. The effectiveness of deep learning architectures in distinguishing between ovarian histopathology pictures that are malignant and those that are not is examined in this study. Three unique deep learning models-Vision Transformer (VIT), MobileNet, and Mobile Former-named Tri-path ensemble architecture are included in the study. Important performance indicators including accuracy, sensitivity, specificity, and AUC are utilized extensively to assess models. The observations reveal significant differences in the models' capabilities, with Mobile-Former standing out as the top structure with remarkable sensitivity and accuracy. In contrast, this surpasses the 98% accuracy attained by Mobile-Former Architecture.