Model Card

Model Overview

obliq-OCR is an Optical Character Recognition (OCR) project designed to extract text from document images.

The default RapidOCR engine uses the PP-OCRv4 detection and recognition weights and the PP-OCR mobile orientation classifier. Their license and model file details are recorded in THIRD_PARTY_NOTICES.

Intended Use

• Extract text from supported document images.
• Process scanned documents.
• Support document digitization workflows.

Limitations

• Recognition accuracy depends on image quality and resolution.
• Blurry, distorted, or low-contrast images may affect extraction.
• Complex document layouts may impact extraction quality.

Evaluation

Refer to BENCHMARK.md for performance benchmarks and evaluation details.

Responsible Use

Extracted text should be reviewed for accuracy before being used in critical workflows.