Multimodal medical volumetric image fusion based on multi-feature in 3-D shearlet transform

Multimodal medical volumetric image fusion (MVIF) plays an important role in medical image processing. This paper presents a novel MVIF method based on multi-feature obtained by a multiscale geometric analysis of the three-dimensional shearlet transform (3DST) with contextual hidden Markov model (CHMM) and type-2 fuzzy logic techniques. First, the preregistered medical volumetric images are decomposed into low frequency parts and high frequency parts by using the 3DST. Next, a choose-max fusion rule with the local energy is performed on the low frequency subbands. Moreover, an effective multi-feature based fusion rule is proposed for integrating the high frequency subbands. The proposed multi-feature consisting of the CHMM statistical characteristics, type-2 fuzzy entropy, gradient and energy of high frequency coefficient is more accurate in representing the source image than single feature, which is favorable to improve the quality of fusion result. Finally, the fused image is obtained by the inverse 3DST with all composite subbands. Experiments on medical volumetric images show that better fusion results can be obtained by the proposed method.