A low complexity wavelet-based blind image quality evaluator

The aim of blind image quality assessment (BIQA) methods is to evaluate the perceptual quality of a distorted image without any prior information regarding its reference image. Although some impressive image quality metrics have been proposed, due to the complexity of the human visual system and the lack of a reference image, designing an image quality metric which accurately predicts human judgments is still a challenging issue. In this paper, a low complexity wavelet-based image quality assessment is proposed. Firstly, the interaction of fine and coarse details of the image, which is extracted by Haar wavelet, is analyzed. In the proposed approach, the joint statistics of two normalized high frequency subbands which indicate coarse and fine structures is utilized for extracting features. Actually, analyzing the relation between image details of different granularities is the main idea of the proposed method. After feature extraction phase, support vector regression (SVR) is adopted in order to provide a quality score. Experimental results show the effectiveness of the proposed low complexity approach.