Analysis and Design of No-Reference Image Quality Assessment
Yuan Tian, Ming Bo Zhu, Ligong Wang · 2008
No-reference (NR) image quality assessment (IQA) seeks to assign quality scores that are consistent with human perception but without an explicit comparison with the reference image. Unfortunately, the field of NR IQA has been largely unexplored. The basic problems in NR IQA are discussed in this paper. And several methods used in NR IQA are explained and analyzed. Finally, an overall NR IQA method is designed. It should be a synthetical index, which applying fuzzy measures and fuzzy integrals, incorporating human vision system (HVS) characteristics and adhering to the philosophy of quantifying quality through blind distortion measurement.