| --- |
| license: cc-by-4.0 |
| pretty_name: SVG Infographic Dataset |
| tags: |
| - svg |
| - infographics |
| - vector-graphics |
| - computer-vision |
| - document-ai |
| - ocr |
| - multimodal-ai |
| - data-visualization |
| size_categories: |
| - n<1K |
| --- |
| |
| # SVG Infographic Dataset |
|
|
| ## Overview |
|
|
| The SVG Infographic Dataset is a curated collection of vector-based infographics intended to support research into how AI systems process and interpret information-rich visual content. |
|
|
| Infographics communicate information through text, graphical elements, layouts, colors, shapes, and spatial relationships. They can represent concepts, explain processes, compare categories, summarize data, and communicate structured information visually. |
|
|
| SVG files preserve visual appearance alongside an XML-based representation of graphic elements. This allows researchers to explore infographics both as rendered images and as structured documents. |
|
|
| This initial release contains **10 sample SVG files** for exploration and experimentation. It is a sample subset, not the complete dataset. |
|
|
| ## Dataset Details |
|
|
| - **Format:** SVG (Scalable Vector Graphics) |
| - **Files:** 10 sample files |
| - **Publisher:** DataCluster Labs |
| - **Release Type:** Initial sample release |
|
|
| ## Why SVG Infographics? |
|
|
| Depending on the individual file, SVG source code may contain: |
|
|
| - Text elements and labels |
| - Vector paths and geometric shapes |
| - Groups and nested elements |
| - Colors, fills, and stroke attributes |
| - Coordinates and positioning information |
| - Relationships between visual components |
| - References to external or embedded assets |
|
|
| These properties support research combining source-level analysis with image-based processing. The presence of individual elements may vary across files. |
|
|
| ## Potential Applications |
|
|
| - Infographic and layout understanding |
| - OCR and text extraction |
| - Multimodal AI and visual question answering |
| - Document intelligence and information extraction |
| - SVG parsing, rendering, and analysis |
| - Chart and diagram understanding |
|
|
| These are potential applications; the dataset is not claimed to be annotated or benchmarked for these tasks. |
|
|
| ## Intended Audience |
|
|
| This dataset may be useful for: |
|
|
| - AI and machine learning researchers |
| - Computer vision engineers |
| - Multimodal model developers |
| - Document AI researchers |
| - OCR and information extraction practitioners |
| - Data visualization researchers |
| - Developers building SVG processing tools |
|
|
| ## Key Features |
|
|
| SVG files can preserve text, shapes, paths, colors, and layout information, enabling analysis of both visual content and underlying document structure. |
|
|
| ## Limitations |
|
|
| This release contains only 10 sample files and may not represent the diversity or scale of the complete collection. |
|
|
| Annotations, ground-truth labels, OCR transcripts, bounding boxes, and standardized dataset splits are not included unless explicitly provided. |
|
|
| ## License |
|
|
| This release is distributed under the **Creative Commons Attribution 4.0 International (CC BY 4.0)** license, provided the publisher has the necessary rights to license all included files under these terms. |
|
|
| Under CC BY 4.0, users may share and adapt the licensed material, including for commercial purposes, subject to the license conditions and appropriate attribution. |
|
|
| The license applies only to material covered by this release. It does not automatically grant rights to unreleased collections, trademarks, or third-party material with separate licensing terms. |
|
|
| License details: https://creativecommons.org/licenses/by/4.0/ |
|
|
| ## Citation |
|
|
| Please use the following citation when referencing this release: |
|
|
| DataCluster Labs. *SVG Infographic Dataset*. Initial sample release, 10 SVG files. |
|
|
| For academic publications or technical reports, include the dataset repository URL and access date. |
|
|
| ## Publisher and Contact |
|
|
| **DataCluster Labs** |
|
|
| Website: https://datacluster.ai/ |
|
|
| For questions about the dataset, additional samples, research collaboration, or access to a larger collection: |
|
|
| Email: sales@datacluster.ai |
|
|
