期刊
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
卷 32, 期 3, 页码 1048-1060出版社
IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
DOI: 10.1109/TCSVT.2021.3073410
关键词
Distortion; Measurement; Task analysis; Image quality; Adaptation models; Data models; Training; No-reference image quality assessment; generalization ability; optimization-based meta-learning; convolutional neural networks
资金
- National Natural Science Foundation of China [61771473, 61991451, 61379143]
- Fundamental Research Funds for the Central Universities [JBF2119, 2021QN1071]
- Key Project of Shaanxi Provincial Department of Education [20JY024]
- Science and Technology Plan of Xi'an [20191122015KYPT011JC013]
- Natural Science Foundation of Jiangsu Province [BK20181354, BK20200649]
- Six Talent Peaks High-level Talents in Jiangsu Province [XYDXX-063]
Recently, there has been great interest in using CNNs for NR-IQA. However, existing metrics in optimizing CNN-based NR-IQA models are limited due to the lack of large training data. In this study, a deep meta-learning based NR-IQA metric is proposed, which achieves high generalization ability and outperforms state-of-the-art methods in both evaluation performance and generalization ability.
Recently, researchers have shown great interest in using convolutional neural networks (CNNs) for no-reference image quality assessment (NR-IQA). Due to the lack of big training data, the efforts of existing metrics in optimizing CNN-based NR-IQA models remain limited. Furthermore, the diversity of distortions in images result in the generalization problem of NR-IQA models when trained with known distortions and tested on unseen distortions, which is an easy task for human. Hence, we propose a NR-IQA metric via deep meta-learning, which is highly generalizable in the face of unseen distortions. The fundamental idea is to learn the meta-knowledge shared by human when evaluating the quality of images with diversified distortions. Specifically, we define NR-IQA of different distortions as a series of tasks and propose a task selection strategy to build two task sets, which are characterized by synthetic to synthetic and synthetic to authentic distortions, respectively. Based on these two task sets, an optimization-based meta-learning is proposed to learn the generalized NR-IQA model, which can be directly used to evaluate the quality of images with unseen distortions. Extensive experiments demonstrate that our NR-IQA metric outperforms the state-of-the-arts in terms of both evaluation performance and generalization ability.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据