Application of Human Visual System in Image Fusion Using Multi Resolution Transforms

Abstract

Image fusion plays a vital role in the field of  Quality of vision. The main objective of the image fusion is gathering important information from multiple images. In this work, image fusion is implemented in transform domain. Identification of important information based on the frequency content is very practical approach.  In the first step, image is decomposed using Discrete Wavelet Transform (DWT). Smoothness is computed in transform domain in the next step. Selection of appropriate band is considered based on the smoothness of the band. Further, Human Visual System (HVS) is also explored for fusion. HVS is used to select the important band in this method. Good texture information is identified using smoothness based fusion technique. It eliminates the noise affected portions. Visually important informationis preserved using HVS based fusion technique. Experiments are performed on multi-focus and medical images. Peak Signal to Noise ratio (PSNR), Mutual Information(MI), Feature Similarity Index(FSIM), and Normalized Cross Correlation (NCC) are used to judge the quality of the fused image. The experimental results show that proposed fusion techniques are good in terms of these fusion metrics.