Multi-focus Image Fusion Algorithms Research Based on Curvelet Transform

Fu, Qiang; Fu-tian, Ren; Le-geng, Chen; Xiao, Zhexin

doi:10.1109/wgec.2009.156

Cited by 3 publications

(1 citation statement)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These algorithms can be mainly divided into two categories: the transformed coefficient fusion method and the sub-block extraction method. The transformed coefficient fusion method proceeds in three steps: 1) transforming source images into another domain by some kind of transforming modes, for instance Laplacian pyramid [1], wavelet [2,3], contourlet [4][5][6], curvelet [7], empirical mode decomposition [8,9], non-negative matrix factorization [10], and so on; 2) integrating the transformed coefficients to get the composite coefficients by diverse schemes, such as the combination of the pixel-level fusion with some aspects of feature-level fusion [11]; 3) inversely transforming the composite coefficients to get the fused image. Different transforming modes have their respective advantages including multi-scale, localization, multidirection, anisotropy, etc.…”

Section: Introductionmentioning

confidence: 99%

Multi-focus image fusion by progressive pixel extraction

Chen

et al. 2011

2011 6th IEEE Conference on Industrial Electronics and Applications

View full text Add to dashboard Cite

In recent years, there has been great progress in the field of multi-focus image fusion. However, existing methods still have their respective defects, and the quality of fused image needs to be improved further. In this paper, a novel fusion method by progressive pixel extraction is proposed. In accordance with the sums of all high-frequency coefficients in a series of windows, the method aims at extracting sharp pixels instead of merging transformed coefficients or extracting sub-blocks as used in other methods, and its extraction process is conducted with progressive strategy at continuously increased window size. As a result, artifacts and block effects are largely reduced in the fused image, and the contradiction between precisely extracting clear edges and extracting larger area is resolved. Practical experimental results demonstrate that the proposed method is far superior to the comparable methods in terms of visual effect and objective evaluation.

show abstract