A Review on Experimental Evaluation Methods for NR-IQA Algorithm
Bo Zhang, Caixia Hao, Ya‐Ru Gao, Xinran Chen, Wenhao Q. Sun, Yanxiang Hu · 2024
Image Quality Assessment (IQA) is a key component of the underlying computer vision community. In order to verify the effectiveness of various objective IQA methods, experimental evaluation is required. In this paper, we mainly summarize the No Reference IQA (NR-IQA) methods in recent 10 years, and focus on the experimental evaluation. Specifically, the rules of subjective IQA and common databases are introduced, we compare and summarize the experimental evaluation methods of NR-IQA based on deep learning, analyze the existing problems, and put forward suggestions and prospects for potential future research.