Image Fusion Using Spectral Selective Correlation In Multiband Coding For Medical Image

Image processing has gained a broader usage for multiple domains in recent time. Advancement in the area of automated diagnosis tool using imaging applications has outcome with advanced methods of information processing and retrieval systems in many real time applications. Preserving detailed information with minimum coefficient representation is a major need in various critical applications such as medical domain. In recent, image fusion has evolved as an approach to develop multi-modality images, which are more informative compared to the conventional medical image representation. Transformed domain multi-modality approach is observed to be a promising solution in recent. Multi-modal fusion is developed based on feature-based technique on transformed information. However, the finer resolutions in transformed domain are observed to be redundant resulting into a processing overhead. The overhead impacts the representation accuracy, and delay decreasing the efficiency of image fusion. To overcome the overhead issue in image fusion approach, this paper outlines a method of spectral correlation in energy level for medical image fusion. Outlined correlative approach process on spectral density and limits the selection of coefficient with minimal loss of information. An adaptive threshold computation for selection of coefficient is proposed. Observations of the proposed approach illustrated a comparatively improved performance in fusion factor, fusion symmetry and entropy measure for medical image processing.