Medical Image Fusion & Classification Of Tumor Cells Using Deep Learning

Abstract

Clinical "image fusion" is a complex course of Consolidating and enlisting different image from Single or numerous
examining methods that can envision the human body for conclusion and tracking down the idea of infections and for
treatment purposes. This interaction is applied to upgrade the nature of the images and furthermore to diminish the
consistency and overt repetitiveness. Through clear sensing and analysis of the fusion, experts may definitely identify
the precise location of the tumor in any part of the brain. Thus, this study presents a deep learning-based fusion-based
categorization model with the following stages: The process involves gathering data from brain MRI and CT scan
images using tools like REMBRANDT and Kaggle, fusing MRI and CT images using variational mode decomposition
(VMD) and local energy maxima (LEM), preprocessing the fused medical image using enhanced CLAHE and nonlocal
mean filter (NLMF) to improve image quality, using a convolutional encoder to extract features, and finally
using a Baysian autoencoder to classify multimodal images. Experimental results show that proposed model
outperforms better than other state-of-art models under various measures (accuracy:0.95, specificity:0.97,
sensitivity:0.97).