Scanned Receipts OCR Dataset with 712 Samples

Voxel51 provides a FiftyOne dataset containing 712 scanned receipt images. The dataset is designed for optical character recognition and information extraction tasks. It was last updated on October 21, 2025.

Use Cases

• Train OCR models based on scanned receipt images.
• Extract structured information like merchant names and totals from receipts.
• Benchmark document understanding systems on real-world receipt data.
• Develop automated expense tracking applications using receipt data.
• Test multimodal data handling with FiftyOne's dataset management tools.

Strengths

• Dataset contains 712 samples, providing a substantial collection for training.
• It is hosted on Hugging Face and integrated with FiftyOne, facilitating easy loading and visualization.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Voxel51

Freshness

Last updated 2025-10-21 18:18:16; freshness should be verified.