SSIM Image Fusion Algorithm Based On Complex Wavelet Transform

With the gradual deepening of the research on image processing technology,image fusion in image processing field has an extensive development prospects,such as medical diagnosis,remote sensing analysis and military detection process.Image fusion is to combine the images of the same scene which are attained with the same or different types of imaging sensors into a single one in order to acquire more correct,more accurate,and more complete descriptions and interpretations of the interested objects.However,Most of the current fusion algorithms are sensitive to the non structural distortions of input images,the SSIM(Structural Similarity Index Measurement)image fusion algorithm based on complex wavelet transform is proposed in this paper.This paper introduces the ideas of the Double Tree Complex wavelet transform into image fusion based on structure similarity.The images is decomposed by Double Tree Complex Wavelet,After a set of suitable fusion rules are applied for decomposed images.In this paper,the fusion method based on image structure similarity is applied to the low frequency coefficients,and high frequency coefficients uses the method of taking largest high frequency coefficients.The experimental results show that the proposed method of fusion image has better visual effect.Moreover,the SSIM fusion algorithm based on complex wavelet transform has a strong robustness to the unstructured distortion of the image,such as smaller rotation,translation and scale.In addition,the fusion image results are comprehensively and objectively evaluated by a large number of experimental data.The main points in this dissertation are as follows:1.The structural similarity based on Complex wavelet transform(CWSSIM)is used to make up for the shortcomings of the spatial domain SSIM index which is very sensitive to the images translation,scale change and rotation.These distortions mentioned above are unstructured distortions,which are not caused by structural changes in the objects of the visual domain.What's more,on the other hand,the CWSSIM index is very sensitive to structural distortions such as JEPG2000 compression,because the structural distortions will result in significant changes in the corresponding phase pattern.This completely solves the shortcomings of the spatial domain SSIM index.Therefore,the CWSSIM index is applied to the image fusion to get an image fusion algorithm which is insensitive to the unstructured distortion of the image and sensitive to the structure distortion.2.As the Double Tree Complex wavelet transform has some advantages such as translation invariance,more directional selectivity,complete reconstruction,and less data redundancy,it has been applied in a series of image processing fields such as image denoising,image enhancement,edge detection,pattern recognition,data compression,texture analysis,and digital watermarking.In this paper,Double Tree Complex wavelet transform is applied to image fusion instead of the Complex wavelet transform,and the translation invariance and more directional selectivity of Double Tree Complex wavelet transform are combined with structural similarity index in order to obtain a fusion algorithm which is suitable for dealing with complex,rich and less unstructured distortion images.