A novel metric for digital image quality assessment using entropy-based image complexity
Pouria Khanzadi, Babak Majidi, Ehsan Akhtarkavan · 2017
Increasing use of image processing algorithms e.g. edge detection, image segmentation, image compression in various commercial applications requires fast and reliable automatic evaluation of the images quality. Image Quality Assessment (IQA) methods mainly use the dissimilarly between the original and the transformed image or the structural elements of the image as the measure for the IQA. In both of these measures the IQA algorithm will not take into account the complexity of the image in calculation of the IQA metric. Using the concept of complexity in calculation of the IQA metric will make it possible to introduce the concept of the observer into the process of assessing the quality of the image. In this paper first a a novel definition for the concept of image complexity based on Shannon entropy is proposed. Then based on this definition a new IQA metric is proposed. The experimental results show that the proposed IQA metric is more sensitive to small changes in the transformed images compared to other proposed IQA metrics.