No-Reference Image Quality Assessment algorithm for Contrast-Distorted Images based on local statistics features

Abstract

Contrast change is a special type of image distortion; it is a very important for visual perception of image quality. Most No-Reference Image Quality Assessment (NR-IQA) metrics are designed for the quality assessment of images distorted by compression, noise and blurring. Few NR-IQA metrics exist for Contrast-Distorted Images (CDI). Existing approaches rely on global statistics to estimate contrast quality. The current No Reference Image Quality Assessment for Contrast-Distorted Images (NR-IQACDI) uses global statistics features. Room for improvement exists, especially for the assessment results using the image database called TID2013 which has poor correlation with Human Visual Perception (HVP); Pearson Correlation Coefficient (PLCC) < 0.7. In this work, instead of relying on the global statistics features, NRIQACDI is presented based on the hypothesis that image distortions may alter the local region statistics (Local patches features). Our experiments are conducted to assess the effect of using local patches features with natural scene statistics (NSS). The experiment results are based on K-fold cross validation with K range from (2 to 10). The statistical tests indicate that the performance using local statistical features are better than that of the NRIQACDI. The use of other statistical features and selection methods should be further investigated to increase the quality of prediction performance. © 2006-2018 Asian Research Publishing Network (ARPN).