USING NOISELET AS A MEASUREMENTMATRIX IN COMPRESSIVE SENSING

The emerging theory of compressive sensing (CS)is an alternative to Shannon/Nyquist sampling theorem speciallyin case of big data size applications. Perfect reconstruction ofundersampled data in CS framework is highly dependent toincoherence of measurement and sparsifying basis matriceswhich is usually fulfilled by selecting a random measurementmatrix. Noiselets as a measurement matrix have a very lowcoherence with wavelets which is the interest of CS, but up tonow they have not been compared with other well knownGaussian and Bernoulli measurement matrices from randomnessview point. So the main purpose of this paper is to introduce theNoiselets and compare them with other measurement matrices intwo point of view; randomness and recovered images.

کلیدواژه ها:

Compressive sensing (CS) ، Noiselets ، Gaussianmeasurement ، Bernoulli measurement ، randomness