An Accurate Diagnosis And Classification Of Brain Tumor Using Transfer Learning In Deep Convolutional Neural Network (DCNN)
Main Article Content
Abstract
Detecting and classifying brain tumors is essential for accurate diagnosis and successful planning of therapy. This study investigates a new method for detecting and categorizing brain tumors by using transfer learning and introducing a modified version of the Inception V3 model. By optimizing a pre-trained Inception V3 network—renowned for its exceptional feature extraction capabilities—the suggested approach uses transfer learning to boost the model's performance. We suggest making alterations to the conventional Inception V3 architecture to enhance its versatility and precision, particularly for the purpose of brain tumor imaging. The enhanced model integrates supplementary convolutional layers and fine-tunes the inception modules, with the objective of capturing more intricate patterns in brain CT data. The dataset used consists of a wide range of brain CT scans, including different categories of cancers as well as non-tumor instances. The experimental findings indicate that the modified Inception V3 model has superior performance compared to conventional and alternative deep learning methods in terms of accuracy, sensitivity, and specificity. The suggested model attains a precision of 92.5%, accompanied by a sensitivity of 91.8% and a specificity of 93.1%. This study emphasizes the effectiveness of transfer learning, together with model improvements, in enhancing the identification and categorization of brain tumors. This contributes to more precise and efficient diagnostic procedures in medical imaging.