arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2408.09371cs.CVcs.AIcs.LG
收藏

检测不可检测之物:结合Kolmogorov-Arnold网络与MLP的AI生成图像检测

Detecting the Undetectable: Combining Kolmogorov-Arnold Networks and MLP for AI-Generated Image Detection

Barisal Information Technology College (BITC)(巴里萨尔信息技术学院(BITC));American International University Bangladesh (AIUB)(孟加拉国美国国际大学(AIUB))

Taharim Rahman Anon, Jakaria Islam Emon

AI总结:

针对先进生成模型导致AI生成图像难检测的问题,提出融合KAN与MLP的混合检测框架,结合语义图像嵌入,在分布外测试中F1分数优于标准MLP,鲁棒性更强。

摘要:

随着人工智能技术发展,先进生成模型使得区分真实图像与AI生成图像的任务愈发复杂。本文提出一种新型检测框架,可稳健识别由前沿生成式AI模型(如DALL-E 3、MidJourney和Stable Diffusion 3)生成的图像。我们构建了一个专门包含这些先进生成器输出图像的综合数据集,作为广泛评估的基础。我们提出了一种将语义图像嵌入与传统多层感知器(MLP)相结合的分类系统,该基线系统可在多种挑战性条件下有效区分真实与AI生成图像。为进一步提升性能,我们提出了一种融合Kolmogorov-Arnold网络(KAN)与MLP的混合架构,该混合模型利用KAN的自适应高分辨率特征变换能力,使系统能够捕捉并分析AI生成图像中常被传统模型忽略的复杂模式。在分布外测试中,我们的模型在三个分布外测试数据集上均持续优于标准MLP,在真实图像与AI生成图像分类任务中展现出优异的性能与鲁棒性,取得了出色的F1分数。

英文摘要:

As artificial intelligence progresses, the task of distinguishing between real and AI-generated images is increasingly complicated by sophisticated generative models. This paper presents a novel detection framework adept at robustly identifying images produced by cutting-edge generative AI models, such as DALL-E 3, MidJourney, and Stable Diffusion 3. We introduce a comprehensive dataset, tailored to include images from these advanced generators, which serves as the foundation for extensive evaluation. we propose a classification system that integrates semantic image embeddings with a traditional Multilayer Perceptron (MLP). This baseline system is designed to effectively differentiate between real and AI-generated images under various challenging conditions. Enhancing this approach, we introduce a hybrid architecture that combines Kolmogorov-Arnold Networks (KAN) with the MLP. This hybrid model leverages the adaptive, high-resolution feature transformation capabilities of KAN, enabling our system to capture and analyze complex patterns in AI-generated images that are typically overlooked by conventional models. In out-of-distribution testing, our proposed model consistently outperformed the standard MLP across three out of distribution test datasets, demonstrating superior performance and robustness in classifying real images from AI-generated images with impressive F1 scores.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑