Image quality assessment (IQA) is traditionally classified intofull-reference (FR) IQA and no-reference (NR) IQA according to whether theoriginal image is required. Although NR-IQA is widely used in practicalapplications, room for improvement still remains because of the lack of thereference image. Inspired by the fact that in many applications, such asparameter selection, a series of distorted images are available, the authorspropose a novel comparison-based image quality assessment (C-IQA) method. Thenew comparison-based framework parallels FR-IQA by requiring two input images,and resembles NR-IQA by not using the original image. As a result, the newcomparison-based approach has more application scenarios than FR-IQA does, andtakes greater advantage of the accessible information than the traditionalsingle-input NR-IQA does. Further, C-IQA is compared with otherstate-of-the-art NR-IQA methods on two widely used IQA databases. Experimentalresults show that C-IQA outperforms the other NR-IQA methods for parameterselection, and the parameter trimming framework combined with C-IQA saves thecomputation of iterative image reconstruction up to 80%.
展开▼
