Ovarian Cancer Detection using Resampling Uncertainty Asynchronous Propagation Penalized Neural Network with Secretary Bird Optimization Algorithm

The detection of ovarian cancer is limited by factors such high tumor heterogeneity, ineffective screening techniques, and late-stage diagnosis brought on by nebulous symptoms. Accuracy is affected by imaging problems such as noise and poor contrast. To increase early detection rates and classification performance, computational models need better feature extraction, segmentation, and optimization methods. Utilizing the TCIA (The Cancer Imaging Archive) dataset, the efficacy of the Resampling Uncertainty Asynchronous Propagation Penalized Neural Network with Secretary Bird Optimization Algorithm (RU-APPNNet-SBOA) for ovarian cancer detection is assessed. First, image quality is improved by the Resampling Cubature Kalman Filter (RCKF). Regions were impacted by the Uncertainty-Aware Decision Transformer (UADT) segments according to variances. Ovarian cancer types are classified via the Asynchronous Propagation Penalized Neural Network (APPNNet), which extracts characteristics. The Secretary Bird Optimization Algorithm (SBOA) examines many impacted areas in the medical imaging to improve categorization accuracy. The TCIA dataset is used in the Python test script, and the results demonstrate that RU-APPNNet-SBOA outperforms current techniques with 99.9% efficiency and 99.8% sensitivity, demonstrating the potential of computer systems to replace manual diagnosis.