Journal article
Jiahao Huang, Tong Liu, Fangyuan Lei, Xiuli Bi, Xiaochen Yuan
IEEE Transactions on Industrial Informatics, 2026 Jun 18, pp. 1-12
IEEE Transactions on Industrial Informatics, 2026 Jun 18, pp. 1-12
Link
Cite
Cite
APA
Click to copy
Huang, J., Liu, T., Lei, F., Bi, X., & Yuan, X. (2026). Omniaggregation Topology Perception Networks for AIGC-Manipulation Detection and Localization. IEEE Transactions on Industrial Informatics, 1–12. https://doi.org/10.1109/TII.2026.3694161
Chicago/Turabian
Click to copy
Huang, Jiahao, Tong Liu, Fangyuan Lei, Xiuli Bi, and Xiaochen Yuan. “Omniaggregation Topology Perception Networks for AIGC-Manipulation Detection and Localization.” IEEE Transactions on Industrial Informatics (June 18, 2026): 1–12.
MLA
Click to copy
Huang, Jiahao, et al. “Omniaggregation Topology Perception Networks for AIGC-Manipulation Detection and Localization.” IEEE Transactions on Industrial Informatics, June 2026, pp. 1–12, doi:10.1109/TII.2026.3694161.
BibTeX Click to copy
@article{huang2026a,
title = {Omniaggregation Topology Perception Networks for AIGC-Manipulation Detection and Localization},
year = {2026},
month = jun,
day = {18},
journal = {IEEE Transactions on Industrial Informatics},
pages = {1-12},
doi = {10.1109/TII.2026.3694161},
author = {Huang, Jiahao and Liu, Tong and Lei, Fangyuan and Bi, Xiuli and Yuan, Xiaochen},
month_numeric = {6}
}
Abstract: As artificial intelligence generated content (AIGC) technologies become increasingly accessible, the risk of malicious image manipulation increases, which poses new challenges to the reliability of industrial visual systems. Compared to traditional tampering, AIGC-edited images often disrupt the semantic topology and structural consistency of visual content, making forgery localization more difficult. To address this, omniaggregation topology perception networks (OTP-Net) is proposed, a novel framework designed to jointly capture pixel-level artifacts and topological inconsistencies. Specifically, a high-frequency discrepancy learning module is introduced to enhance fine-grained manipulation traces, while spatial and channel topology perception branches are designed to capture multiscale topological anomalies through graph-based modeling. Extensive experiments demonstrate that OTP-Net achieves competitive performance in CelebA-HQ and AutoSplice datasets, while showing notable gains in the average pixel-level F1 and intersection over union by up to 3.86% and 4.42% in challenging composite attacks.