Identification and Categorization of Brain Tumors Using HOG Feature Descriptor

The precise diagnosis of brain tumors is aided by early detection, which is essential in medical diagnostics. Medical diagnosis using Magnetic Resonance Imaging (MRI) is a viable alternative, but manual inspection is difficult owing to the enormous volumes of data and fine nuances. Radiologists depend on the ability to recognize cancers in MRI scans, which is where computer-aided diagnostics come into play. One of the most common and fatal malignancies, brain tumors can shorten a person's typical lifetime. The survival of patients with brain tumors is improved by early identification. Medical imaging can be difficult to use to detect tumors because of their dimensions, form, and place. It is essential to comprehend these specifics in order to understand and categorize malignancies. In medical imaging, machine learning techniques may distinguish between tumor and healthy pictures. Preprocessing, feature extraction, and classification make up a suggested strategy for improving the simplicity of identifying brain cancers. Raw data is preprocessed to remove noise, and HOG feature description is used to extract features. The system's accuracy is assessed using a variety of machine learning classifiers, such as Support Vector Machine (SVM), Gradient Boost, K Nearest Neighbor (KNN), XG Boost, and Logistic Regression. The classifier with the best accuracy, Extreme Gradient Boosting (XG Boost), achieves nearer to 92%.