Multifocus image fusion based on Uniform Discrete Curvelet Transform

Abstract

A novel multi-focus image fusion method based on Uniform Discrete Curvelet Transform (UDCT) is proposed to overcome conventional muti-scale analysis image fusion shortcomings, such as high data redundancy ratio, complicated structure and poor performance, etc. First, UDCT is applied to the multi-focus images and subband coefficients of multi-scales and multi-directions are obtained. Then, different fusion rules are used for high-low frequency coefficients respectively. i.e. high frequency coefficients are fused by rules of local energy, and low frequency coefficients are fused by rules of local average gradient. Finally, Inverse Uniform Discrete Curvelet Transform (IUDCT) is applied to the new high-low frequency coefficients, and the fused result image is obtained. Simulation results show that the proposed UDCT images fusion method is superior to other related multi-scale analysis image fusion methods in subjective visual effect and objective evaluation parameters.