- Problem
- Bengali-script documents are poorly served by mainstream OCR tooling, which is built and tuned around Latin-script text — a real barrier to digitizing Bengali-language archives.
- Approach
- An end-to-end OCR pipeline for Bengali script digitization, built as a B.Sc. thesis project at Daffodil International University under Dr. Md Zahid Hasan.
- Result
- A working end-to-end pipeline converting scanned Bengali PDFs to editable text, distinct from the CT reconstruction research line and demonstrative of broader engineering range.