SPECT and MR Image Fusion Scheme based on Hybrid Activity Measures with Consistency Verification inNon-Subsampled Contourlet Transform Domain

Abstract

Integration of SPECT and MR images provide composite information in a single fused image where both functional and anatomical features can be viewed simultaneously.Medical practitioners need these multimodal images to achieve fast and accurate diagnosis of human health issues. In this paper, Hybrid activity measures are proposed to integrate multi-sensor images in Non-subsampled contourlet transform (NSCT) domain. The NSCT is used to decompose source images into low frequency (approximation) and high frequency(details) subbands. The proposed method includes different hybridfusion rules for these complementary low and high frequency subbands. For the selection of low frequency subband information, a hybrid fusion scheme based on local energy and local entropy is used. The high frequency subbands are merged by using a weighted sum of Laplacian coefficients and weighted local energy.Finally, spatial domain fused imageis obtained via Inverse NSCT. In this work, a comparison is accomplished with recent fusion methods by choosing multimodal brain images. The experimental results reveal the effectiveness of the proposed hybrid activity measures with consistency verification fusion in terms of the image quality and quantitative assessment.

Keywords:Nonsubsampled contourlet transform, Single Positron Emission Computed Tomographyimage, Magnetic resonance Imaging, Local energy, Entropy, Laplacian.