FakeShield: Explainable Image Forgery Detection and Localization via Multi-modal Large Language Models
Paper • 2410.02761 • Published • 2
FakeShield Model Card
Model details
Model type: FakeShield is a multi-modal framework for explainable image forgery detection and localization (IFDL). It integrates multi-modal large language models (MLLMs) to analyze manipulated images, generate tampered region masks, and provide explanations. The model includes weights for both DTE-FDM and MFLM components.
The model repository includes weights for both DTE-FDM and MFLM components.
Model date: fakeshield-v1-22b was trained in September 2024.
Project page: https://zhipeixu.github.io/projects/FakeShield/
License
FakeShield is licensed under the Apache 2.0 License.
Where to send questions or comments about the model:
https://github.com/zhipeixu/FakeShield/issues
Intended use
Primary intended uses: The primary use of FakeShield is used for research on image forgery detection and localization, enhancing transparency in media forensics.
Primary intended users: The primary intended users of the model are researchers, forensic analysts, journalists, and AI practitioners in media forensics and misinformation detection.
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support