I. Introduction
Being able to automatically evaluate image quality, as perceived by human beings, plays an important role in many image processing and computer vision applications. This naturally raises the demand for objective image quality assessment (IQA) metrics. While the interests in objective IQA have been growing at an accelerated pace over the past decades and the latest progress in this field is encouraging, there also has been an abundant menu of unresolved problems left for further study [1]. Among the unresolved problems, quality analysis of tone-mapped images (TMIs) is one of the most challenging problems while the effort made to this area remains quite limited.