Compressive sensing based microarray image acquisition

This paper implements Orthogonal Matching Pursuit (OMP) algorithm for reconstruction of Microarray Images based on the compressive sensing paradigm. Gaussian and Bernoulli random patterns are used to capture the scene. A Monte Carlo simulation is performed to calculate the peak signal to noise ratio, relative error and universal quality index of the red and green channels of the image independently. Since images are not sparse but rather compressible, the paper seeks to reconstruct approximately 90 percent of the energy using Discrete Cosine Transform (DCT) as the basis. This paper successfully proposes the use of post processing for quality improvement rather than increase in measurements. Post processing using median filter can account for a reduction of 200 samples per block. The results obtained show that, both the sensing matrices tested are equally good with Bernoulli pattern having the advantage of being sparse.