Detecting AI-Generated Forgeries via Iterative Manifold Deviation Amplification

Jiangling Zhang; Shuxuan Gao; Bofan Liu; Siqiang Feng; Jirui Huang; Yaxiong Chen; Ziyu Chen

cs.CVFeb 21, 2026

Detecting AI-Generated Forgeries via Iterative Manifold Deviation Amplification

Authors: Jiangling Zhang, Shuxuan Gao, Bofan Liu, Siqiang Feng, Jirui Huang, Yaxiong Chen, Ziyu Chen

Organizations: 1Wuhan University of Technology

Abstract

The proliferation of highly realistic AI-generated images poses critical challenges for digital forensics, demanding precise pixel-level localization of manipulated regions. Existing methods predominantly learn discriminative patterns of specific forgeries and often struggle with novel manipulations as editing techniques continue to evolve. We propose the Iterative Forgery Amplifier Network (IFA-Net), which shifts from learning "what is fake" to modeling "what is real". Grounded in the principle that all manipulations deviate from the natural image manifold, IFA-Net leverages a frozen Masked Autoencoder (MAE) pretrained on real images as a universal realness prior. Our framework operates through a two-stage closed-loop process: an initial Dual-Stream Segmentation Network (DSSN) fuses the original image with MAE reconstruction residuals for coarse localization, followed by a Task-Adaptive Prior Injection (TAPI) module that converts this coarse prediction into guiding prompts to steer the MAE decoder and amplify reconstruction failures in suspicious regions for precise refinement. Extensive experiments on four diffusion-based inpainting benchmarks show that IFA-Net achieves an average improvement of 6.5% in IoU and 8.1% in F1-score over the second-best method, while demonstrating strong generalization to traditional manipulation types.

Figures & tables

Figure 1: Conceptual comparison of our IFA-Net against existing methods. Existing methods often leverage Diffusion Models for reconstruction in a single-pass, open-loop process that lacks task-adaptive guidance.Our method is grounded in a different prior: a Masked Autoencoder (MAE) pretrained on real images. We introduce a two-stage, closed-loop framework where the coarse prediction from Stage 1 guides a second, targeted reconstruction in Stage 2 for significant anomaly amplification.

Figure 2: Overall architecture of IFA-Net. The framework operates in two stages while sharing the same Dual-Stream Segmentation Network (DSSN). Stage 1 (blue arrows): The input image $x$ is processed by a fully frozen MAE (Encoder & Decoder) to produce an initial residual map $x_{\text{rec}}$ . The DSSN then fuses $x$ and $x_{\text{rec}}$ to predict a coarse mask $M_{\text{crs}}$ , supervised by $\mathcal{L}_{\text{crs}}$ . Stage 2 (red arrows): The coarse mask $M_{\text{crs}}$ is encoded into Task-Adaptive Prompts ( $T_{p}$ ), which modulate features from the frozen MAE Encoder via a FiLM layer. The modulated features are passed to a trainable MAE Decoder to generate an amplified residual map, which is then used by the shared DSSN to predict the final refined mask $M_{\text{ref}}$ , supervised by $\mathcal{L}_{\text{ref}}$ .

Figure 3: Visualization of anomaly amplification. We compare reconstructed images and residual maps from Stage 1 and Stage 2. In Stage 1, the frozen MAE can already produce coarse forgery cues—although weak and noisy, they indicate that the MAE indeed encodes a realness prior. In Stage 2, prompt-guided reconstruction forces the model to fail more strongly on forged regions, further amplifying these forgery cues and yielding a much cleaner and more salient residual map that better supports the final segmentation.

	TruFor		PSCC-Net		MVSS-Net		SPAN		IML-ViT		MaskCLIP		DcDsDiff		Ours
Dataset	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1	IoU	F1
GIT Benchmarks
OpenSDI	0.369	0.426	0.297	0.368	0.298	0.349	0.284	0.349	0.325	0.373	0.427	0.494	0.385	0.445	0.487	0.620
GIT10K	0.770	0.815	0.392	0.754	0.453	0.518	0.137	0.257	0.882	0.859	0.694	0.765	0.801	0.864	0.928	0.963
CocoGlide	0.481	0.552	0.769	0.869	0.419	0.533	0.276	0.411	0.291	0.440	0.783	0.878	0.878	0.935	0.887	0.940
Inpaint32K	0.711	0.762	0.116	0.207	0.496	0.578	0.456	0.541	0.703	0.775	0.390	0.480	0.790	0.852	0.811	0.896

Table 1: Quantitative results on GIT and TT benchmarks. Comparison of the proposed method with recent state-of-the-art approaches. Values are reported as IoU / F1. Best and second-best results for IoU and F1 are independently shown in bold and underline , respectively.

Figure 4: Qualitative comparison on OpenSDID. Examples are selected from five diffusion generators (SD1.5, SD2.1, SD3, SDXL, and Flux.1). IFA-Net produces cleaner and more accurate localization masks that align closely with the ground truth, whereas other methods often yield incomplete or fragmented detections.

	Components			GIT-AVG		TT-AVG
Index	DSSN	TAPI	AD	IoU	F1	IoU	F1
I	✗	✗	✗	0.670	0.770	0.459	0.608
II	✓	✗	✗	0.684	0.782	0.483	0.625
III	✓	✓	✗	0.745	0.823	0.554	0.670
IV	✓	✓	✓	0.778	0.855	0.586	0.708

Table 2: Ablation study of IFA-Net. Performance impact of each component on diffusion-based (GIT-AVG) and traditional tampering (TT-AVG) benchmarks. Index I represents a single-stream baseline that concatenates the original image and residual map as input. AD denotes the Adaptive Decoder. Best results are highlighted in bold .

Figure 5: Robustness analysis under Gaussian blur and Jpeg compression.

Explore similar work

Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization
Jun 1, 2026Yiming Wang, Baiqi Wu, Qingming Li +3Image Forgery Detection Digital Image Forensics
Zhejiang University, Hangzhou, China.
Progressive Decision-Making for Localizing Open-Ended AI-Generated Image Forgeries
Jul 31, 2026Jingyi Hou, Xiaoxia Chen, Leyu Zhou +2Image Segmentation Image Forgery Localization
School of Artificial Intelligence, University of Science and Technology Beijing, Beijing 100083, China · Institute of Artificial Intelligence, University of Science and Technology Beijing, Beijing 100083, China · Key Laboratory of Intelligent Bionic Unmanned Systems, Ministry of Education, University of Science and Technology Beijing, Beijing 100083, China
Structured Local Differential Modeling for AI-Generated Image Detection
Aug 13, 2026Jiazhen Yang, Ruijin Jin, Junjun Zheng +3Image Forgery Detection Digital Image Forensics
Zhejiang University · Taobao & Tmall Group of Alibaba · Zhejiang University of Technology

Jun 1, 2026J/K move · Enter open · S save

Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization

Yiming Wang, Baiqi Wu, Qingming Li, Jiahao Chen+2

Zhejiang University, Hangzhou, China.

Image Forgery Detection Digital Image Forensics Image Forgery Localization Anomaly Detection

Open paper AskPDF arXiv

Recent advancements in generative AI have led to image editing models capable of producing realistic forgeries that evade traditional image forgery localization methods, as these approaches depend on physical noise absent in synthetic data. To address this challenge, we theoretically demonstrate that the diffusion process inherently suppresses local high-frequency variance, creating a statistical energy gap that is distinguishable from the natural entropy of optical imaging. Guided by this insight, we propose FLAME, a unified framework that utilizes a LAD map to capture these intrinsic anomalies, coupled with a parameter-efficient adapter for SAM to achieve precise, pixel-level forgery localization. Furthermore, to bridge the lag between forensic benchmarks and evolving generative models, we introduce EditStream, an automated pipeline for continuous, instruction-based training data synthesis. Extensive experiments demonstrate that FLAME establishes a new state-of-the-art, significantly outperforming previous methods on AI-generated forgery datasets while effectively generalizing to unseen generative architectures. Our code is available at https://github.com/phoenixnir/FLAME.

Abstract

Explore similar work

Order within Chaos: Capturing Intrinsic Energy Anomalies for AI-Manipulated Image Forgery Localization

Progressive Decision-Making for Localizing Open-Ended AI-Generated Image Forgeries

Structured Local Differential Modeling for AI-Generated Image Detection