No-reference image quality assessment algorithm for contrast-distorted images enhanced by using directional contrast feature in curvelet domain

Reduced-reference Image Quality Metric for Contrast-changed images (RIQMC) and No-Reference Quality metric for Contrast-Distorted Images (NR-IQACDI) are the state-of-the-art IQA for Contrast-Distorted Images (CDI). Nevertheless, there is room for improvement especially for the assessment results using image database called TID2013 and CSIQ. Most of the existing No-Reference Image Quality Assessment Algorithm (NR-IQA) metrics designed for CDI use features in spatial domain. In the current work, we pursue to compliment it with feature in Curvelet domain which is powerful in capturing multiscale and multidirectional information in an image. Indeed, the Directional Contrast (DC) is captured in the Curvelet domain of CDI by decomposing the image into several directional subbands across multiple scales using curvelet transform. Due to the fact that high-frequency subband consists of many directional information, the directional contrast of each directional subband coefficient is generated as feature vector. Finally a Support Vector Regressor (SVR) is used to predict the image quality score. Experiments are conducted to assess the effect of adding DC feature in the Curvelet domain. The experimental results based on i-fold cross validation with K ranging from 2 to 10 and statistical test indicate that the performance of NRIQACDI can be improved by adding DC feature in the Curvelet domain.