OmniAID: Decoupling Semantic and Artifacts for Universal AI-Generated Image Detection in the Wild

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Guo, Yuncheng, Ye, Junyan, Zhang, Chenjue, Kang, Hengrui, Fu, Haohuan, He, Conghui, Li, Weijia
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866918529299644416
author Guo, Yuncheng
Ye, Junyan
Zhang, Chenjue
Kang, Hengrui
Fu, Haohuan
He, Conghui
Li, Weijia
author_facet Guo, Yuncheng
Ye, Junyan
Zhang, Chenjue
Kang, Hengrui
Fu, Haohuan
He, Conghui
Li, Weijia
contents A truly universal AI-Generated Image (AIGI) detector must simultaneously generalize across diverse generative models and varied semantic content. Current methods learn a single, entangled forgery representation, conflating content-dependent flaws with content-agnostic artifacts, and are further constrained by outdated benchmarks. We propose OmniAID, a novel framework centered on a decoupled Mixture-of-Experts (MoE) architecture that separates: (1) semantic flaws across distinct content domains via Routable Specialized Semantic Experts, and (2) content-agnostic universal artifacts from content-dependent flaws via a Fixed Universal Artifact Expert. A two-stage training strategy first specializes experts independently with domain-specific hard-sampling, then trains a lightweight gating network for effective input routing. By explicitly decoupling "what is generated" (content-specific flaws) from "how it is generated" (universal artifacts), OmniAID achieves robust generalization. We also introduce Mirage, a large-scale, contemporary dataset comprising a modern training set and a challenging test set. Extensive experiments demonstrate that OmniAID surpasses existing detectors, establishing a new standard for AIGI detection against modern, in-the-wild threats. Code is available at https://github.com/yunncheng/OmniAID.
format Preprint
id arxiv_https___arxiv_org_abs_2511_08423
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OmniAID: Decoupling Semantic and Artifacts for Universal AI-Generated Image Detection in the Wild
Guo, Yuncheng
Ye, Junyan
Zhang, Chenjue
Kang, Hengrui
Fu, Haohuan
He, Conghui
Li, Weijia
Computer Vision and Pattern Recognition
A truly universal AI-Generated Image (AIGI) detector must simultaneously generalize across diverse generative models and varied semantic content. Current methods learn a single, entangled forgery representation, conflating content-dependent flaws with content-agnostic artifacts, and are further constrained by outdated benchmarks. We propose OmniAID, a novel framework centered on a decoupled Mixture-of-Experts (MoE) architecture that separates: (1) semantic flaws across distinct content domains via Routable Specialized Semantic Experts, and (2) content-agnostic universal artifacts from content-dependent flaws via a Fixed Universal Artifact Expert. A two-stage training strategy first specializes experts independently with domain-specific hard-sampling, then trains a lightweight gating network for effective input routing. By explicitly decoupling "what is generated" (content-specific flaws) from "how it is generated" (universal artifacts), OmniAID achieves robust generalization. We also introduce Mirage, a large-scale, contemporary dataset comprising a modern training set and a challenging test set. Extensive experiments demonstrate that OmniAID surpasses existing detectors, establishing a new standard for AIGI detection against modern, in-the-wild threats. Code is available at https://github.com/yunncheng/OmniAID.
title OmniAID: Decoupling Semantic and Artifacts for Universal AI-Generated Image Detection in the Wild
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2511.08423