Objective:Doctors can get different details from different medical images because of the different imaging theory of medical instruments and application

image fusion become a key means in clinic diagnose and medical research.Methods:This paper adopted an image fusion method based on non-separable wavelet frame transform.Firstly

having finished the registration

medical image was decomposed with non-separable wavelet decomposition

and into low and high frequency sub-image.According to the different characteristics of high frequency and low frequency components

the paper acquire different fusion rule to fuse the images

then through the inverse wavelet transform

we can acquire the fused image.In addition

the paper introduced the entropy

average gradient

uniformity of fusion to evaluate the fused image.Results:The results showed that this method could enhance the detail and brightness information of fused image

which is superior to traditional wavelet transform.Conclusion:It is an effective method of medical image fusion based on the non-separable wavelet framework image fusion algorithm

which could improve the accuracy and validity of disease diagnosis.