REVIEW ON DEEP LEARNING APPROACH FOR BRAIN TUMOR GLIOMA ANALYSIS

Abstract

Brain tumor diagnosis has evolved as a very critical need in current medical diagnosis. Early diagnosis of tumor detection is an important need for the primitive treatment of brain tumor patient increasing the survival rate of patient. MRI diagnosis of brain tumor for cancer treatment is a large processing due to volumetric content of scan sample. The processing of clinical data is large and consumes a high processing time. Hence, the need of early diagnosis and proper segmentation of brain tumor region is in need. This paper outlines a review on the developments of MRI sample processing for early diagnosis for brain tumor glioma diagnosis using deep learning approach. The advantage of learning capability and finer processing efficiency has gained an advantage in MRI image processing, which enable a better processing efficiency and accuracy in early diagnosis. Deep learning approach has shown a benefit of image coding based on selective features and state of art processing in diagnosis. The evaluation objective of the MRI sample processing has shown a better accuracy than the comparative existing approaches.  The recent trends, the advantages and limitation of the existing approach for MRI diagnosis is outlined.