文献类型：期刊文献

英文题名：Study of Subjective and Objective Naturalness Assessment of AI-Generated Images

作者：Chen, Zijian[1];Sun, Wei[1];Wu, Haoning[2];Zhang, Zicheng[1];Jia, Jun[1];Huang, Ru[3];Min, Xiongkuo[1];Zhai, Guangtao[1];Zhang, Wenjun[1]

机构：[1]Shanghai Jiao Tong Univ, Inst Image Commun & Informat Proc, Shanghai 200240, Peoples R China;[2]Nanyang Technol Univ NTU, S Lab, Singapore 639798, Singapore;[3]East China Univ Sci & Technol, Sch Informat Sci & Engn, Shanghai 200237, Peoples R China

年份：2025

卷号：35

期号：4

起止页码：3573

外文期刊名：IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY

收录：;EI(收录号：20244917491087);WOS:【SCI-EXPANDED(收录号:WOS:001459809200017)】；

基金：This work was supported in part by China Postdoctoral Science Foundation (CPSF) under Grant 2023TQ0212 and Grant 2023M742298; in part by the Postdoctoral Fellowship Program of CPSF under Grant GZC20231618; in part by Shanghai Pujiang Program under Grant 22PJ1407400; and in part by the National Natural Science Foundation of China under Grant 62271312, Grant 62301316, Grant 62101325,Grant 62101326, and Grant 62132006.

语种：英文

外文关键词：AI-generated images; image naturalness assess-ment; image quality assessment; dataset; dataset; dataset

摘要：The proliferation of Artificial Intelligence-Generated Images (AIGIs) has greatly expanded the Image Naturalness Assessment (INA) problem. Different from early definitions that mainly focus on tone-mapped images with limited distortions (e.g., exposure, contrast, and color reproduction), INA on AI-generated images is especially challenging as it owns more diverse contents and could be affected by factors from multiple perspectives, including low-level technical distortions and high-level rationality distortions. In this paper, we take the first step to benchmark and assess the visual naturalness of AI-generated images. First, we construct the AI-Generated Image Naturalness (AGIN) dataset by conducting a large-scale subjective study to collect human opinions on the overall naturalness as well as perceptions from the technical quality and rationality perspectives. AGIN verifies several insights for the first time that naturalness is universally and disparately affected by both technical and rational distortions, while its manifestations vary with different generation tasks. Second, to automatically assess the naturalness of AIGIs that align with human opinions, we propose the Joint Objective Image Naturalness evaluaTor (JOINT). Specifically, JOINT imitates human reasoning in naturalness evaluation by jointly learning technical and rationality features with several specific designs to guide model behavior from respective perspectives. Experiments demonstrate that JOINT significantly outperforms existing methods for providing more subjectively consistent results on naturalness assessment. The dataset can be accessed at https://github.com/zijianchen98/AGIN.