Integration of Machine Learning and IoT for Assisting Medical Experts in Brain Tumor Diagnosis

The irregular development of tissue in the brain or spine is known as a brain tumor. A tumor is of two types that are malignant or benign. A brain tumor can be differentiated between primary as well as secondary. When its diagnosis is done using a traditional approach where segmentation is not as skilled as when done with help of ML or deep learning and it’s a time taking process. IoT helps in making treatment hassle-free for the patient as well as provides better communication for both. A lot of IoT-based devices help in self-assessment. A better solution compared to the manual approach is that of the Machine learning-based approach in terms of quality and labor. Various Machine learning processes for detection could be used. Segmentation, and classification through deep learning seems a better approach than machine learning. Comparison of different methods of feature extraction, segmentation method, and classification method using Machine learning and then by deep learning is done separately in the tables. An optimized algorithm is really important for the best result. So here an optimized way is used which includes the use of IoT along with CWOA and ANN for a precise result. CWOA is explained which is derived from WOA along with pseudocode. Post getting the scans a sensor is used to convert images into electrical signals and store them in the server after that in Pre-processing, a median filter can be used to reduce noise. For image segmentation, Otus’s method is used for thresholding. Various features are extracted using texture, statistics and for feature selection, CWOA has been used and for classification, ANN is used with CWOA. Later a comparison is done among all the models and for day to day implementation the accurate result is handled to the server which the patients can use with help of a mobile IoT based application.