Early Brain Tumor Prediction Using Hybrid Optimized Fuzzy Clustering-Active Contour Segmentation Based Heuristic Deep Learning Model

Biomedical technology is crucial in identifying and managing serious health conditions that offer significant risks to life. The deadliest common diseases has been identified recently. The brain tumor detection and classification system is available so that it can be diagnosed at early stages. To help with the detection of brain tumors, radiologists and physicians greatly benefit from the use of an automated tumor detection system. Currently, physicians manually examine the patient’s MR images of their brain to ascertain the brain tumor’s location and dimensions. This is deemed to be exceedingly time-consuming and leads to incorrect tumour detection. As technology has advanced recently, medical professionals believe that integrating machine learning and MRI picture data could be a useful way to recognize and categorize brain cancers. Brain tumor patients are categorized using machine learning techniques on a large scale. However, the limitations of conventional works include the inability to handle huge datasets, erroneous location identification, incorrectly classified outcomes, and more sophisticated algorithm design. Therefore, an enhanced deep learning methodology is established to proficiently identify brain tumor ailments. The designed model gathers MRI data in standard source. The image data are preprocessed because they contain noise and irrelevant elements that may affect the prediction process as well as consume more time. The preprocessing techniques are resizing, MPR net, and adaptive histogram contrast normalization, which are employed in the proposed model. Subsequently, the preprocessed data is segmented using a hybrid technique based on improved fuzzy C means active contour segmentation, which efficiently identifies the disease portion. After that, segmented images are fed using an optimized OZnet classifier thus effectively predicting tumor or non tumor. The performance metrics attained from the proposed model are 96.86 \(\%\) of accuracy, 93.58 \(\%\) of precision and 93.5 \(\%\) of recall. The proposed model’s obtained values are better than those of the current methods. Thus, the proposed optimized deep learning approach for effectively detecting brain tumor disease with higher accuracy.