Forbin Dataset: Historical Photographs with Archival Metadata and Annotations

The Forbin Dataset is a large-scale collection of historical photographs taken or collected by Victor Forbin (1868–1947). It provides COCO-style segmentation polygon annotations alongside archival metadata such as Box ID, description, notes, and dates. The dataset is hosted on HuggingFace by author mchelali and was last updated in December 2025.

Use Cases

• Train image segmentation models based on the COCO-style polygon annotations.
• Analyze historical visual content and its evolution based on the collection's temporal span.
• Enrich archival records by linking image content with the provided descriptive metadata.
• Develop tools for exploring and previewing historical image collections using the included HTML/JS interface.

Strengths

• Includes COCO-style segmentation annotations, enabling machine learning tasks.
• Contains archival metadata fields like Box ID, description, notes, and dates.
• Features a lightweight HTML/JS explorer interface for previewing images and annotations.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The temporal and geographic coverage of the photographs is not explicitly detailed.

Provenance

Source

Photographs taken or collected by Victor Forbin.

Collection Method

Collection and annotation process not detailed.

Time Range

Covers the period of Victor Forbin's life (1868–1947), but specific photo dates are 'when available'.

Freshness

Last updated 2025-12-09 08:49:50; freshness should be verified.