No-reference image quality assessment in complex-shearlet domain

The field of image quality measure (IQM) is growing rapidly in recent years. In particular, there was a significant progress in no-reference (NR) IQM methods. Natural scenes have certain statistical properties which vary in the presence of distortion. The statistical changes represent the loss of naturalness and can be efficiently quantified using shearlet transformation of images. In this paper, a general-purpose NR IQM approach is proposed based on the statistical characteristics of natural images in shearlet domain. The method utilizes a set of distortion-sensitive features extracted from statistical properties of shearlet coefficients. Phase and amplitude of an image contain important perceptual information; therefore, a complex version of the shearlet transform is employed to take advantage of phase and amplitude features in quality estimation. In quality prediction step, the features are used to train image classification and quality prediction models using a support vector machine. The experimental results show that the proposed NR IQM is highly correlated with subjective assessment and outperforms several full-reference and state-of-art NR IQMs.