Evaluation of image quality using dual-tree complex wavelet transform and compressive sensing

Proposed is a new image quality metric using a dual-tree complex wavelet transform and compressive sensing (CS) that represents a sparse signal with a small number of measurements. Structural distortions in an image change both magnitude and phase in the complex domain. Sparse complex wavelet (CW) coefficients obtained from CW transform (CWT) can also reflect the structural distortion in an image. These CW coefficients can be obtained using the dual-tree CWT. If CW coefficients are measured via CS, the measurements impose the information of the image. Thus, the measurements are used for assessing the image quality by comparing measurements of the reference and distorted images. Experiments with the LIVE database show the effectiveness of the proposed method.