Diagnostic support system for detecting and grading brain tumor

Abstract

newline With rapid technology development, healthcare sector has received increasing attention for digitizing and automating diagnostic procedures to save time and to support in different diagnostic opportunities. Medical imaging is receiving importance with improved image processing techniques, towards retrieval of relevant information from medical images with the potential of providing information for accurate diagnosis, monitoring drug therapy reactions and illness management of patients. Despite the challenges seen medical imaging has proven worth in diagnosing many diseases. In this research work, diagnostic support systems that utilize three conceptually distinct disciplines like radiology, histopathology and genomics data have been developed for brain tumor diagnosis. Early and accurate diagnosis of brain tumor using AI and machine learning models, that assist radiologists and pathologists in grading tumors for appropriate treatment of the patients to lengthen their lifespan has been the main objective of this research. The first part of the proposed work is related to improve the diagnosis of tumor with fused radiological images obtained from various modalities. Multimodal radiological images are decomposed into high frequency sub bands and low frequency sub bands which are then fused and segmented using Improved Intersecting Cortical Model (IICM) which is proposed to obtain series of correlated feature vectors of the fused image. Gray-Level Co-occurrence Matrix (GLCM) is used for significant features extraction. The extracted features are classified using Long Short-Term Memory (LSTM) into Low Grade Glioma (LGG) and High Grade Glioma (HGG). newline newline