An Image Fusion Algorithm Based on Directionlet Transform

An image fusion algorithm based on Directionlet transform was proposed to improve the performance of image fusion.Firstly,two registered original images were decomposed into cosets separately by the generator matrix and the sub-images corresponding to each coset were obtained.Secondly,the low frequency and high frequency sub-band coefficients were obtained by subtraction of each sub-image from another,and the singular features such as edge and texture were included in high frequency sub-band coefficients.Then the selection of low frequency coefficients was performed with direct average combination and for high frequency sub-band coefficients with stronger edge-texture information of sub-areas.Finally,the fused image was obtained using inverse Directionlet transform.Experimental results of multi-focus image fusion indicate that the algorithm is much better in fusing some image features such as the edge in subjective visual,thus can better keeps details of source images,and in objective evaluation,it also has better performance than wavelet tranform and other image fusion algorithms by comparing parameters of entropy,average gradient,standard deviation and mutual information.