Fusion of ultrasound and magnetic resonance images for endometriosis diagnosis: a non-parametric approach

A fusion method was recently proposed for ultrasound and magnetic resonance images for endometriosis diagnosis. This method combined the advantages of each modality, i.e., the good contrast and signal to noise ratio of the MR image and the good spatial resolution of the US image. The method was based on an inverse problem, performing a super resolution of the MR image and a denoising of the US image. A polynomial function was introduced to model the relationships between the gray levels of the MR and US images. This paper studies the potential interest of replacing this polynomial function by a non-parametric transformation built using the theory of reproducing kernel Hilbert spaces. Simulations conducted on a phantom and synthetic data allow the performance of the resulting fusion method to be appreciated.