Deep Learning Techniques for Brain Lesion Classification Using Various MRI (from 2010 to 2022): Review and Challenges

Abstract

Brain tumors are conditions brought on by the development of aberrant brain cells. They are classified into non-cancerous (benign) and cancerous (malignant). The morbidity and mortality of brain tumors are challenging to determine. A study in the United Kingdom disclosed that around 15 out of every 100 individuals with brain cancer could survive for ten or more years after being diagnosed. The remedial maneuvers of the brain tumors depend upon the kind of brain tumor, degree of cellular abnormality, location of Cancer in the brain, and other variables. The treatment decision needs assistance from the Deep learning algorithms using magnetic resonance imaging (M.R.I.) data to diagnose brain tumors due to the high dimensionalities of the remedial maneuvers. MRI is a scanning technique that uses strong radio waves and strong magnetic fields to generate detailed images of the body's interior. The study employed deep learning models to detect the tumor region in brain M.R.I. scans, including a Convolutional Neural Network model. The proposed processes involved dataset modification and preprocessing, detection, identification, and classification via CNN. Data mining techniques were utilized to uncover significant relationships and patterns from the data, resulting in successful early brain lesion identification and classification using deep learning approaches.

Received: 4 September 2023 | Revised: 7 November 2023 | Accepted:  19 December 2023

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support this work are available upon reasonable request to the corresponding author. The dataset analyzed for another part of the study can be found in the BraTS 2020.