Dataset · Hugging Face

MohamedDhouib1/TDoc-2.8M

Name: TDoc-2.8M
Creator: MohamedDhouib1
License: mit
Keywords: task_categories:image-segmentation, language:en, language:fr, language:zh, license:mit, size_categories:1M<n<10M, arxiv:2602.17322, region:us, document-image-analysis, document-forensics, tamper-detection, forgery-detection, synthetic-data, multilingual, computer-vision

TDoc-2.8M TDoc-2.8M is a large-scale multilingual dataset for tampered text detection and localization in document images. It contains approximately 2.8 million tampered document images from multiple languages. The dominant languages are English, French, and Chinese. The full dataset size is about 7.44 TB. Resources 📄 Paper: arXiv:2602.17322 💻 Code and generation pipeline: Mohamed-Dhouib/TDOC Dataset Format The dataset is distributed as… See the full description on the dataset page: https://huggingface.co/datasets/MohamedDhouib1/TDoc-2.8M.

View on Hugging FaceReport an issue

Author / organization	MohamedDhouib1
License (as declared)	mit
Downloads	15.9K (as reported by Hugging Face)
Likes	2
Created	18 February 2026
Last updated at source	3 October 2026

document-image-analysisdocument-forensicstamper-detectionforgery-detectionsynthetic-datamultilingualcomputer-visiontampered-text-detectiondocument-forgery-localizationdocument-tampering-localizationimage-manipulation-detectioncontrastive-learningdocument-tampering-detectiondocument-forgery-detection

Embed this entry

<iframe src="https://silytics.si/embed/item/e43d972a-2e17-46d9-a083-7c9f0c8e37e5" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 14 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.