Magnetic resonance and computed tomography image fusion using saliency map and cross bilateral filter

Abstract

An image fusion technique is suggested for magnetic resonance and computed tomography source imaging using saliency map and cross bilateral filter. Since human anatomy consists of many complicated structures, clear demarcated visibility is essential for physicians to deeply analyze the suspicious areas. The source snapshots are disintegrated into various base members and detailed member segments which are melded and consolidated to acquire the resultant fused image. The simulations are performed on variety of images obtained from standard Harvard image database. The proposed technique is able to preserve boundaries (of soft tissues and bones) in the fused image. The fused image obtained using proposed technique is compared with existing techniques using state-of-the-art objective and fusion metrics. The quantitative analysis suggests that the proposed scheme mostly outperforms state-of-the-art schemes in terms of edge, saliency and structural information.