Adaptive PET/CT fusion using empirical wavelet transform

Abstract

Empirical Wavelet Transform (EWT) is an adaptive signal decomposition technique in which the wavelet basis is constructed based on the information contained in the signal instead of a fixed basis as in standard Wavelet Transform (WT). Its adaptive nature enables EWT in many image processing applications like image denoising, image compression, etc. In this paper, a new adaptive image fusion algorithm is proposed for combining CT and PET images using EWT. EWT first decomposes both the images into approximate and detailed components using the adaptive filters that are constructed according to the content of an image by estimating the frequency boundaries. Then, the corresponding approximate and detailed components of CT and PET images are combined by using appropriate fusion rules. An adaptive EWT image fusion, a newly proposed method, is compared with standard WT fusion using the image quality metrics, image fusion metrics and error metrics. The quantitative analysis proved that the newly proposed method results in better quality than the standard WT method.