A Hybrid Multimodal Medical Image Fusion Technique for CT and MRI Brain Images

Estimating the type, size, location and spread of a brain tumor is vital in the diagnosis and treatment of brain cancer. Fused CT and MRI brain images assist in faster detection and diagnosis of brain tumors. They provide superior results in comparison to individual CT or MRI images. Multiscale transforms (MSTs) are widely used in fusing multimodal images like CT and MRI. However, they have a few drawbacks like reduced contrast, poor edge detection, redundancy and high computation time. This article describes how MSTs coupled with sparse representation (SR) aims to overcome the drawbacks. Non-Subsampled Contourlet Transform (NSCT) is widely used on MSTs for fusing multifocal images. Therefore, a novel technique using NSCT and SR is proposed for better quality fused CT and MRI images. The experimental results show superior performance.