Feature selection algorithm for no-reference image quality assessment using natural scene statistics

Abstract

Images play an essential part in our daily lives and the performance of various imaging applications isdependent on the user’s quality of experience. No-reference image quality assessment (NR-IQA) has gained importanceto assess the perceived quality, without using any prior information of the nondistorted version of the image. DifferentNR-IQA techniques that utilize natural scene statistics classify the distortion type based on groups of features and thenthese features are used for estimating the image quality score. However, every type of distortion has a different impacton certain sets of features. In this paper, a new feature selection algorithm is proposed for distortion identificationbased image verity and integration evaluation that selects distinct feature groups for each distortion type. The selectionprocedure is based on the contribution of each feature on the Spearman rank order correlation constant (SROCC) score.Only those feature groups are used in the prediction model that have majority features with SROCC score greater thanmean SROCC score of all the features. The proposed feature selection algorithm for NR-IQA shows better performancein comparison to state-of-the-art NR-IQA techniques and other feature selection algorithms when evaluated on threecommonly used databases.