Automated Prediction of Brain Tumor Recurrence Using an Adaptive Sail Fish Based Inception V3

The most severe cancer that causes death is brain tumor (BT). Even after standard treatment, few of the huge grade BT easily recurrent. While managing patient results, a critical component is brain tumor recurrence prediction (BTRP) that are more aggressive in nature. The traditional approaches like molecular markers, histopathology, and imaging demonstrated significant perceptions, nevertheless, it failed to achieve early and accurate BTRP. Hence, the novel method is required to predict the early risk of BT recurrence thereby the patient's survival time potentially increased. Before and after surgery, this study proposes a novel optimized Inception V3 model for BTRP. The bilateral filtering and min-max normalizations remove the noise and normalize the input images during pre-processing. In segmentation, develop an Adaptive Coarse-to-Fine Attention Refinement (AC-FAR) to solve the issues in unbalanced data. Deep Autoencoder to extract features and progression-free survival area prediction for constructing preoperative validatable and stable from the peritumoral edema and tumor region. The Adaptive Sailfish Optimization (ASFO) and Inception V3 to identify the early recurrence of BT patients at high risk thereby validating and training the system effectively under the Python implementation platform. The proposed work results show superiority over the state-of-art methods.