A Performance Analysis of Measurement Matrices used in Compressed Sensing Techniques

Main Article Content

IshaniMishra, et. al.


 Compressed sensing technique requires three major stages: sparse representation, measurement, and sparse recovery. It performs with sparse representation of real world signals. In compressed sensing based signal acquisition, the input signal, measurement matrix and a measurement vector are required. The compressive measurements are discovered via the multiplication of random measurement matrix and input signal. The count of measurements taken here is not exactly the signal length. Hence, it utilizes a measurement matrix to test just the parts that best addresses the sparse signal. The decision of the measurement matrix influences the accomplishment of the process of sparse recovery. Consequently, the structure of a suitable measurement matrix is a significant interaction in compressive sensing. Absurd years, a few measurement matrices have been determined. Thusly, a brief survey of these measurement matrices and a correlation of their exhibitions is emphatically required. This paper gives an outline on compressed sensing featuring the measurement process. Then, it classifies the measurement matrices and compares the performances of such matrices. The performance comparison of measurement matrices is carried out using few evaluation metrics such as sparse reconstruction error, processing time and covariance.

Article Details