Contrastive Distortion‐level Learning‐based No‐reference Image‐quality Assessment

Xuekai Wei,Jing Li,Mingliang Zhou,Xianmin Wang
DOI: https://doi.org/10.1002/int.22965
IF: 8.993
2022-01-01
International Journal of Intelligent Systems
Abstract:A contrastive distortion‐level learning‐based no‐reference image‐quality assessment (NR‐IQA) framework is proposed in this study to further effectively model various distortion types with the same or different distortion levels. The proposed method aims to improve the prediction accuracy of NR‐IQA. The proposed method consists of three parts: multiscale distortion‐level representation learning, single‐image NR‐IQA, and a representation affinity module, which can reduce NR‐IQA computational complexity while maintaining a low‐distortion representation of high‐distortion inputs. The proposed NR‐IQA method aims to extract distributional features of samples in real distorted images and predict ambiguity based on distortion‐level learning. Experimental results show that by comparing on many NR‐IQA data sets the proposed method can outperform state‐of‐the‐art methods.
What problem does this paper attempt to address?