Comparison on Different Random Basis Generator of a Single-Pixel Camera

Compressive sensing is a signal processing technique that takes advantage of signal sparseness in some domain. To use compressive sensing, a domain in which the signal is represented as a few significant coefficients should be defined. If the proper domain is identified as a set of basis vectors, the coefficients are the projections of the signal on the basis vectors. This is typically a transformation from the original signal space to a lower dimensional signal space. To reverse the transformation, we need to solve an underdetermined linear system. Natural signals such as images and videos are sparse. Therefore, many researches apply compressive sensing as image compression method. Single-pixel camera is one of the interesting topics. It sequentially measures the voltages from the photodiode as the transformed coefficients. The sensing matrix is implemented by a digital micro-mirror device, and can be easily configured using a pseudo random number generator. In this paper, we performed a few experiments based on the algorithms of single-pixel camera. We are interested in the effects of different random basis. Hence, sensing matrices constructed by different random number generators are experimented and discussed.