STM Journals Peer-reviewed research from Consortium e-Learning Network

Journal of Image Processing & Pattern Recognition Progress

A Novel Approach of Medical Image Fusion using Wavelet Transforms

  1. Surya Prasada Rao Borra
  2. Rajesh K Panakala
  3. Rajesh Kumar P

Abstract

Image processing applications have been growing rapidly in real world. The term fusion means an approach to extract the useful information from several modalities. Image fusion (IF) is used to integrate the complementary information obtained from multisensor, multiview and/or multitemporal and get an image of more information and the quality of which cannot be achieved from any individual image. Different fusion algorithms are useful in many applications like medical diagnosis of CT, MRI and/or PET images. Particularly in brain medical imaging, fusing various modalities like MRI and CT images into a single image with in depth anatomical information and high spectral information is needed  in clinical diagnosis. This process not only helps for better diagnosis, it also helps to reduce the storage space of medical images. Image fusion can be possible at three different levels viz.,  pixel level, feature level and decision level. This paper discusses pixel level based fusion algorithms like Averaging fusion, Discrete Wavelet Transform (DWT) based averaging fusion and DWT maximum approximation algorithms. The algorithms are implemented using LabView and simulation results are obtained for various fusion techniques applied on two different data sets of MRI and CT scan images were compared using the quality assessment parameters Mean Squared Error(MSE), Peak Signal to Noise Ratio (PSNR), Entropy  and Structural Similarity Index (SSIM). It is observed that DWT maximum approximation fusion technique yields better result compared to other implemented algorithms.

Keywords

Support