计算机技术与发展2025，Vol.35Issue(11)：1-11,11.DOI:10.20165/j.cnki.ISSN1673-629X.2025.0152
人工智能生成图像检测综述
Review of AI-Generated Image Detection
摘要
Abstract
Advanced generative models have been able to produce realistic images that are indistinguishable to the naked eye,raising a series of security and ethical issues,including the dissemination of false information,and the study of AI-generated image detection techniques is a necessary means of coping with these issues.Although some review studies related to the detection of generated images have appeared in China,they are mainly oriented to the detection of generated adversarial network images,and there is a lack of review studies oriented to the general detection technology of generated images.In order to fill this gap and help researchers engaged in AI generated image detection to understand the current technological progress and existing problems,firstly,the background and significance of AI generated image detection are elaborated.Secondly,representative and influential deep learning models and commercial platforms in the field of image generation are introduced,and then representative methods for detecting generated images are systematically sorted out.Finally,the problems faced by AI-generated image detection are discussed,and the future research focus and development trend are outlooked,providing ideas for further improving the AI-generated image detection methods.关键词
生成图像检测/图像生成模型/通用检测器/检测范式/跨模型Key words
generated image detection/image generation model/generic detectors/detection paradigms/cross-model分类
计算机与自动化引用本文复制引用
杨皓然,马瑞强,王钢,崔旭,郭亚楠..人工智能生成图像检测综述[J].计算机技术与发展,2025,35(11):1-11,11.基金项目
国家自然科学基金(62472237) （62472237）
内蒙古自然科学基金(RZ2300001581) （RZ2300001581）
