Implementation of image fusion Algorithm for improved brain tumor Detection using low power Architecture

Abstract

Medical image processing has developed as one of the critical factors newlinein regular clinical applications such as disease diagnosis and treatment planning newlineOwing to the technical limitations the quality of medical images is usually newlineUnsatisfactory degrading the accuracy of human interpretation and further newlinemedical image analysis thereby requiring the quality of these images to be newlineenhanced One approach to enhance the image quality is by image fusion which newlineimproves the image quality by combining the corresponding information from newlinemultimodal images into a single modified image This resulting image called as newlinefused image provides an accurate sketch of the human body thereby improving newlinethe accuracy in diagnosis Image fusion is a process by which multiple input newlineimages of the same scene are combined into a single fused image The fused newlineimage retains the full content information and the important features from each newlineof the original images newlineIn this research a fusion rule which aims at combining the pixel newlinevalues of the MRI and PET brain images of the unique patient using Discrete newlineWavelet Transform DWT while preserving the contrast has been proposed newlineCertain fusion rules are applied for merging of coefficients at various scales newlineusing Mamdani fuzzy rules Then an inverse DWT is applied over the fuzzy rule newlineapplied image to generate the final fused image The fused image is validated newlineusing the quantitative measures such as Peak Signal to Noise Ratio PSNR newlineMean Square Error MSE and fusion latency newline newline