Revolutionizing Medical Imaging for Ovarian Cancer Detection with TRANS U-Net

Despite improvements in therapy and research for ovarian cancer, the hardest part is still finding tumors in their early stages, which is when women die. To aid with diagnosis, CT scans of the ovarian area were processed using deep learning algorithms. Prior to tumor segmentation using the UNet and TRANS U-NET models, the images underwent a battery of pre-processing procedures. After that, the cases were divided into two groups: benign tumors and malignant tumours. An example of a deep learning model used for classification was VGG16. Investigating the selection of several hyperparameters for the segmentation of the ovary and ovarian follicles, this work delves into the architecture of the U-Net. This research introduces a new method for ovarian cancer detection using U-net and TRANS U-Net segmentation. A TRANS U-Net design, which takes the classic 2D U-Net and expands it into three dimensions in order to deal with volumetric input, is at the heart of our segmentation method. When it comes to 3D medical scans for ovarian cancer, the TRANS U-Net is the way to go because of how well it performs in segmentation of biomedical images tasks. Using this approach, medical imaging apps with strong 3D segmentation structures may identify ovarian cancer much more accurately. Predicting ovarian cancer and its subtypes has been much improved with the deployment of TRANS UNet model, reaching an impressive 96% accuracy. In comparison, the 94% accuracy achieved by UNet network Architecture is surpassed by this.