A Novel Approach to Identify Pituitary Tumor Using an Optimal CNN Model with Deep Learning

The technical paper introduces a novel approach to identifying pituitary tumors using an Optimized Convolutional Neural Network (OCNN) model with deep learning. Pituitary tumors rank among the most prevalent variety of brain tumors, and accurate identification is critical for proper diagnosis and treatment. The suggested method makes use of a deep learning framework that combines an ideal feature extraction method with a pre-trained CNN model. The pulled out features are then used to train the CNN model to classify pituitary tumors accurately. The effectiveness of the suggested method is assessed using a publicly accessible dataset, revealing that it surpasses the performance of current leading techniques. The suggested method possesses the capacity to act as a valuable asset for making clinical decisions in the field of neuroimaging, particularly for the accurate identification of pituitary tumors. The results of the experiment indicate that our suggested model successfully detects and classifies tumors with high accuracy, sensitivity, and specificity. Our model specifically achieves a 99.55% overall accuracy, 99% sensitivity, 99% specificity and 99% F1 score. These outcomes show that our suggested CNN design for brain tumor identification and classification performs better than several state-of-the-art techniques.