Magnetic resonance and computed tomography image fusion using bidimensional empirical mode decomposition

Image Fusion has been widely used for medical images to improve diagnosis accuracy and time by providing medical personnel with more comprehensive picture of the patient condition, where a single modality cannot provide. In this work, we explore image fusion using Empirical Mode Decomposition (EMD) for medical imaging purposes. In particular, we use Bidimensional Empirical Mode Decomposition (BEMD) to analyze Magnetic Resonance (MRI) and Computed Tomography (CT) Images and fuse the generated Bidimensional Intrinsic Mode Functions (BIMFs) using simple fusion rules. BEMD is particularly useful for medical images since the fused images are, in general, anatomically consistent. Thus, BEMD is more likely to yield homogeneous BIMFs, which in turn are easy to fuse computationally. Results of BEMD-based fusion are reported and compared with two other fusion techniques: Curvelet Fusion and Wavelet Fusion. Performance of BEMD is evaluated using perceived quality as well as using three popular image fusion quality metrics; namely, Peak Signal-to-noise Ratio (PSNR), Structure Similarity Index Metric (SSIM), and Mutual Information parameter (MI).