JIN Xin, ZHOU Bin, ZOU Dongqing, LI Xiaodong, SUN Hongbo, WU Le
Image quality assessment (IQA) is to measure the perceived image degradation while image aesthetic quality assessment (IAQA) is to automatically give a review of aesthetics of an image. The composition, color, illumination, depth of field, blurness, sharpness, etc. are analyzed by means of computer to analog the aesthetic perception and cognition of human beings. IAQA is a novel interdisciplinary field of computer aesthetics and computer vision, psychology, and virtual reality. IAQA can be used for photography, movie and television, ecommerce, fashion design, graphic design, document typesetting, industry design, automobile industry, architectural design, beauty makeup, etc. In this article, we give a brief history of IAQA. Then we review the methods based on hand-crafted aesthetic feature design, the methods based on deep aesthetic feature learning, new topics in IAQA, and dataset collection for IAQA. Finally, we give the future directions of IAQA.
