Multi-Focus Image Fusion Using Discrete Cosine Harmonic Wavelet Transform

Iman M. G. Alwan

Abstract: Multi-focus image fusion algorithms are used to combine a collection of different focuses images that represent the same scene, to produce a sharper one. In this paper, a transform domain multi-focus image fusion algorithm is proposed. It depends on using Discrete Cosine Harmonic Wavelet Transform and pixel-based rule of fusion. Region entropy is used to fuse coefficients of approximation subbands, while maximum rule with consistency verification is utilized in fusion of detail subbands of the Discrete Cosine Harmonic Wavelet Transform domain. The fused resultant image is obtained by applying inverse Discrete Cosine Harmonic Wavelet Transform to the combined coefficients. The proposed method is compared with several transform domain based fusion algorithms, and it shows better performance or similar performance under several objective criteria.

Keywords: multi-focus image fusion, discrete cosine harmonic wavelet transforms DCHWT, entropy region, consistency verification