No-reference image quality metrics are of fundamental interest as they can be embedded in practical applications. The main goal of this paper is to define a new selection process of attributes in no-reference learning-based image quality algorithms. To perform this selection, attributes of seven well known no-reference image quality algorithms are analyzed and compared with respect to degradations present into the image. To assess the performance of these algorithms, the Spearman Rank Ordered Correlation Coefficient (SROCC) is computed between the predicted values and the MOS of three public databases. In addition, an hypothesis test is conducted to evaluate the statistical significance of performance of each tested algorithm.
Christophe Charrier, Abdelhakim Saadane, Christine Fernandez-Maloigne, "Dimension reduction-based attributes selection in no-reference learning-based image quality algorithms" in Proc. IS&T Int’l. Symp. on Electronic Imaging: Image Quality and System Performance XIV, 2017, pp 15 - 20, https://doi.org/10.2352/ISSN.2470-1173.2017.12.IQSP-219