No reference image quality assessment: Feature fusion using relevance vector machine
Besma Sadou, Atidel Lahoulou, Toufik Bouden · 2017
In this letter, a novel no reference image quality metric is developed, a set of ten features are extracted from each distorted image, then Relevance Vector Machine algorithm (RVM) is utilized to learn the mapping between the combined features and human opinion scores, experiments are conducted on the LIVE databases. The performance of the proposed metric is compared with some existing NR metrics on LIVE database. Pearson correlation coefficient (PCC) and Spearman Rank Order Correlation Coefficient (SROCC) are used to evaluate the performance of the proposed metric.