A Novel Medical Image Fusion Algorithm based on Wavelet Packet Transform and Visual Saliency Modelling

Authors

  • Jie Yang, Li Hua Ding, Dong Yang, Xiao Zhang, Sun Xinghua

Abstract

The rapid development of information science has put forward the trend of combining the traditional medical processing methods with the state-of-the-art computer science techniques. Under this background, in this paper, we conduct research on the new medical image fusion algorithm based on the wavelet packet transform and visual saliency modelling. Because of the image data in local area is relatively simple, so local model compared with the global model is more reasonable. In local learning, on the basis of the panchromatic image and the fusion image of global regression error is expressed as the form of the graph Laplacian, its essence is the use of topical learning the fusion image to keep the panchromatic image manifold structure. At the same time in order to maintain the core nature of the multispectral image, by the scale of the image space, said the building of the scale of the relationship between image and multispectral image fusion. In addition in the signal or image analysis, sometimes need to signal in the time domain and frequency domain features or image in the airspace and combine characteristics of frequency domain analysis. Therefore, we integrate the wavelet packet transform and visual saliency to propose our algorithm. The experiment shows that the method could extract the primary features with saliency and fuse the image well compared with other latest methodologies. 

Published

2020-04-30

Issue

Section

Articles