Clinical Integration of Positron Emission Tomography (PET) and Magnetic Resonance Imaging (MRI) for Enhanced Medical Image Fusion
Abstract
Medical image fusion enhances image reliability by integrating key data from multiple sources. This paper introduces a fusion method using NSST (Non-Sampled Shearlet Transform) in the transform domain. Entropy analysis, based on PCNN (Pulse Coupled Neural Network), is employed to optimize fusion. The proposed algorithm prioritizes highinformation bands for fusion, resulting in superior image quality. NSST and PCNN are applied to fuse MRI and PET images, ensuring comprehensive data integration.
Keywords
Download Options
Introduction
Advancements in medical imaging equipment offer diverse modalities such as X-ray for bone display, Computed Tomography (CT) for hard tissue visualization, Magnetic Resonance Imaging (MRI) for soft tissue depiction, and Positron Emission Tomography for physiological and pathological insights. Medical image fusion enhances diagnostic accuracy by consolidating information from different imaging modalities, facilitating easier diagnosis. Moreover, image fusion finds applications in various fields including space research, defense, and remote sensing. Transforming images from the time domain to the frequency domain enables efficient analysis. Wavelet Transform, a multi-resolution image decomposition tool, separates images into detailed and approximation coefficients, capturing different image features through frequency sub-bands. Discrete Wavelet Transform (DWT) efficiently handles 1-D singularity, offering improved spectral content, albeit with limited directionalities and time invariance issues. To address these, Shifted Wavelet Transform (SWT) or un-decimated DWT is employed. Various fusion algorithms are developed based on these transforms, with the proposed edge and energy method outperforming average and maximum techniques. This method leverages high-edge information and energy from decomposed bands, yielding superior fusion results.
1.1 DWT
The wavelet transform evaluates signal compatibility with wavelets, yielding higher transform values for matched signalwavelet pairs and lower values otherwise. It's instrumental in discerning regional characteristics within images. Discrete Wavelet Transform (DWT) decomposes images into bands via decimation, achieved through successive 1-D transforms along rows and columns. This decomposition yields sub-bands half the size of the original signal, implemented through scaling. DWT facilitates multi-resolution analysis, with the LL band resembling a spatial image while higher bands capture diverse frequencies. The decomposition and reconstruction of the image is done by using ‘db2’.
Conclusion
The fusion of images using PCNN and NSST occurs in the transform domain, with LL and higher bands selected based on input from two images. Entropy, derived from parameters like mean, median, and pixel gradient, guides the selection process. This method can be further enhanced by decomposing images into multiple levels. Experimental evaluations utilizing NSST and PCNN demonstrate superior performance compared to alternative techniques.