study and investigation of computer aided diagnosis tool for detection of brain tumour

Abstract

CBTRUS statistics account that brain tumours are a significant cause for cancer-related newlinemorbidity and mortality, with a five-year survival of 35% patients. The impetus for improved newlineprognosis would be at early tumour detection and characterisation, enabling planning for the newlineappropriate and timely treatment. newlineEarly clinical diagnosis of a brain tumour is always a challenging task as patients present newlinewith a plethora of signs and symptoms. The working diagnosis confirmed with the gold newlinestandard medical imaging modality being magnetic resonance imaging (MRI). newlineThe parameters requisite for appropriate brain tumour classification and prognosis are sites, newlinesize, shape, and characteristics of the tumour. Manual detection of these brain tumour is time- newlineconsuming as they would require the radiologist to read though multiple MR scan images. newlineThis tedious manual process of image review could delay the speed of diagnosis and is newlinepotentially human error-prone. newlineThere has been a significant improvement in the MRI machine process refinement and hence, newlinethe accuracy of image capture. The review of literature proved that gaps exist at image newlineenhancement and classification process. In the era of artificial intelligence, there is an infinite newlineneed for the development of automatic computer-aided diagnosis tool, which could answer newlinethis crucial need of the hour. newlineVarious methods have been implemented to enhance and boost the performance of newlineclassification to identify tumour state. The research study shows that different pre-processing newlinemechanisms have different effect and impact on the performance of the classification newlineprocedure. The proposed research work considers the analytical methodology and presents a newlineunified work that considers the input brain MRI which is subjected to multiple image newlineenhancement operations to obtain a superior class of enhancement on an input image. newlineThe enhance output subject to segmentation, converts to the binarised image. The multi-level newlineDWT decomposition mechanism obtains decomposition coefficients. newline