Imagetranslation Visual Terminology Dataset

Imagetranslation.org is an open multilingual and multimodal dataset that links domain-specific terminology with annotated illustrations. Each term is assigned a label on an image, translated into multiple languages, and optionally described with a short definition or usage context.

This dataset is designed to help train and evaluate language models, multimodal systems, and educational tools by providing visual grounding for technical concepts.

✨ Key Features

Multilingual support (term_en, term_ru, term_uz, term_cn, term_de, etc.)
Visual identification using numbered arrows
Domain categorization (e.g., Mechanical Engineering, Oil and Gas)
Structured, extendable JSON format
Designed for integration with real-world educational and AI systems

📁 Dataset Structure

🖼️ Example Illustration

Each entry in the dataset contains the following fields:

Field	Description
`image_file`	Name of the image file (e.g., `track_schematic_diagram.png`)
`image_caption_en`	Image caption in English (optional)
`image_caption_ru`	Image caption in Russian (optional)
`category`	Domain or subject area of the image
`terms`	A list of labeled terms shown on the image

Each term includes:

Field	Description
`label_number`	The number shown on the image label or arrow
`term_*`	The term in one or more supported languages (e.g., `term_en`, `term_ru`)
`context_term_*`	Short definition or usage context in a specific language

You may omit any language-specific field (term_cn, context_term_de, etc.) if it is not yet filled. These fields are expandable and can be added later.

🛠 Editing and Expanding the Dataset

This dataset can be expanded using the visual interface at https://imagetranslation.org.

Through the interface, you will be able to:

Upload a clean image (without any labels or text)
Add arrows and numbers using built-in tools
Add and edit terms in multiple languages
Fill in definitions and contexts

🔧 Note: The site is currently under active development.
At the moment, not all fields can be added and JSON export is not yet available.
However, core functionality works — including drawing arrows, numbering, and editing terms.

🤝 How to Contribute or Collaborate

We welcome collaboration!

If you would like to:

contribute your language skills,
help with development or refinement of the platform,
financially support the project,
or hire our team to prepare a custom visual dataset for your AI application —

please contact the author directly.

👤 Author: Pyak Pavel
📍 Location: Tashkent
📧 Email: [email protected]

📜 License

This demo dataset is released under the Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) license.
You are free to use, share, and adapt this dataset for non-commercial purposes with proper attribution.

🔒 Commercial use of this dataset or access to the full version is only available upon request.
Please contact the author for licensing and collaboration opportunities.

🔗 CC BY-NC 4.0 License Overview