The challenge
Searchable Balochi digital text is limited. OCR could help make scanned written material more accessible to text-processing tools.
Approach
- 01Explore document collection and transcription workflows.
- 02Identify possible evaluation datasets and quality measures.
- 03Iterate on recognition pipelines once suitable data is available.
Documented outcome
Research direction in progress; no published model-quality metric is claimed.
Interpretation & limitations
This page describes ongoing work rather than a completed or evaluated public model.