A Comparative Analysis on Image Fusion Algorithms based on Compressive Sensing

This paper is about study of comparative analysis of Spatial and Transform domain fusion techniques under Compressive Sensing or Compressive Sampling principle. The compressive measurements of two source images are obtained using star shaped sampling pattern and fuse the measurements. The output image is reconstructed from 25% of samples using Minimum Total Variation method with equality constraints and with reduced computational time. Finally, for different fusion techniques under Compressive Sensing are performed and compared. Multi focus and Multi modal images are used for simulation and no prior knowledge of source images is required for reconstruction. Based on fusion evaluation metric with reference and without reference image conclude that in spatial domain, simple average & principal component analysis and in transform domain, DCTav and Laplacian Pyramid are performed well.