Infinity-Doc-55K: 55,000 Scanned Documents with Structural Annotations

Infinity-Doc-55K contains 55,000 real-world and synthetic scanned documents for full-text parsing, published by infly in 2025. The collection features layout variations and structural annotations across financial, medical, and academic report domains.

Use Cases

• Training layout analysis models using structural annotations to identify document segments
• OCR benchmarking across financial and medical report types
• Fine-tuning multimodal models for document-based question answering

Strengths

• 55,000 document records
• Structural annotations for layout parsing
• Multi-domain coverage including medical and financial reports

Limitations

• Specific schema and column definitions are not provided in the metadata
• Potential distribution mismatch between synthetic and real-world samples

Provenance

Source

infly (Arxiv 2506.03197)

Collection Method

Hybrid of real-world collection and synthetic generation

Freshness

Last updated November 2025