African Medical Records: Nigerian Handwritten Clinical Notes for OCR

African Medical Records (AMR) is a large-scale, growing dataset of paired handwritten medical notes and their digital counterparts. It is designed as a benchmark for optical character recognition, medical text extraction, and healthcare data analysis. The dataset was created by Nigeria-Health-data-OCR-pipeline and was last updated on 2026-04-17.

Use Cases

• Train optical character recognition models based on handwritten medical notes.
• Extract structured clinical information based on paired handwritten and digital text.
• Model healthcare insights based on data from Nigerian clinical records.
• Benchmark multimodal text extraction systems based on a dedicated dataset.

Strengths

• Dataset is described as large-scale and growing.
• Contains paired data of handwritten notes and digital text, enabling supervised learning.
• Explicitly designed as a benchmark dataset for OCR and clinical text extraction.

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

Nigeria-Health-data-OCR-pipeline on Hugging Face

Collection Method

Likely collected from clinical settings, but specific gathering method is not detailed.

Freshness

Last updated 2026-04-17 23:10:40; freshness should be verified.

Geography

Nigeria