InfoBayAI/Bounding_Box_with_Annotation_Dataset

15K+ Bounding Box Annotated Images Dataset

Dataset Description

The 15K+ Bounding Box Annotated Images Dataset is a general-purpose computer vision dataset containing 15,000 images with bounding box annotations and object class labels. It covers diverse real-world visual content, including storefronts, street signs, documents, billboards, sidewalk advertisements, and restaurant menus etc.

The images include multilingual visual content, including Hindi, Chinese, Thai, Hebrew, Korean, Arabic and Japanese. The dataset is designed to support object detection, visual recognition, scene understanding, and computer vision model development involving commercial environments, public signage, advertisements, and document-related imagery.

By combining image data with bounding box annotations and class labels, the dataset provides visual examples that can help AI developers train and evaluate models for detecting and localizing objects in complex scenes.

Dataset Details

• Dataset Size: 15,000 images
• Data Type: Annotated images
• Image Format: PNG
• Annotation Format: JSON
• Annotation Type: Bounding boxes and object class labels
• Image Content: Storefronts, street signs, documents, billboards, sidewalk advertisements, and restaurant menus etc.
• Languages Represented: Hindi, Chinese, Thai, Hebrew, Korean, Arabic and Japanese languages
• License: cc-by-4.0

Supported Tasks

• Object Detection: Detect and localize objects using bounding box annotations.
• Visual Recognition: Identify and classify objects appearing in images.
• Scene Understanding: Analyze storefronts, streetscapes, advertisements, and commercial environments etc.
• Sign and Advertisement Detection: Locate street signs, billboards, and sidewalk advertisements etc.
• Document and Menu Detection: Identify documents and restaurant menus in visual scenes.
• Multilingual Visual AI: Support computer vision systems processing images containing Hindi, Chinese, Thai, Hebrew, Korean, Arabic Japanese and other language content.

Data Product Features

• Image Identification: Image IDs and filenames used to identify individual images.
• Bounding Box Annotations: Coordinates defining the location of objects within images.
• Object Class Labels: Labels identifying categories such as storefronts, street signs, documents, billboards, sidewalk advertisements, and restaurant menus.
• Coordinate System: The coordinate convention used to represent bounding boxes, subject to confirmation from the annotation files.
• Image Metadata: Available attributes such as image dimensions, file format, and filename.
• Annotation Metadata: Additional annotation attributes, such as annotation IDs, confidence scores, or annotation method, where available.

Intended Uses

The dataset may be useful for:

• Training object detection models.
• Developing computer vision systems for retail and commercial environments.
• Detecting storefronts, street signs, billboards, and sidewalk advertisements.
• Identifying documents and restaurant menus in images.
• Building visual search and image understanding applications.
• Supporting research into multilingual scene understanding.
• Evaluating object localization and recognition systems, where suitable evaluation splits and ground-truth annotations are available.

Geographic coverage

Global

License

This dataset is distributed under the Creative Commons Attribution 4.0 International (CC BY 4.0) License.

Data Creation

Procured through formal agreements and generated in the ordinary course of business.

Considerations

This dataset is provided for research and educational purposes only. It contains only sample data. For access to the full dataset and enterprise licensing options, please visit our website InfoBay.AI or contact us directly.

-Ph: (+91) 9236398619
-Email: datareq@infobay.ai

---