Deep Learning Techniques for Brain Tumor Diagnosis: A Review

Deep Learning techniques have remarkably contributed to the advancement of medical image analysis by strengthening prediction accuracy, lead to proper drafting and diagnosis. Automated medical diagnosis using deep learning techniques help doctors, radiologists and clinical experts in the early detection and diagnosis of diseases. The conventional method for detecting the presence of lesions is more time consuming and labour-intensive. In this paper, we focus on reviewing various deep learning-based techniques used in the early identification of the diagnosis of brain tumors. These diagnosis tasks include feature extraction, segmentation, grading, classification, and prediction. This work carried out a detailed review of state-of-the-art innovations performed on each task related to brain tumor images. We summarized and analysed significant contributions over recent years and investigated their extensive advantages, limitations and dataset specification used in the experiments. Eventually, we addressed the ongoing challenges and future research propositions for practitioners in the domain.