pdf-ocr-parser

I came across a problem, where I have to read some info written in kannada language (one of the widely spoken language of southern state of India) from a scanned pdf or a picture based pdf.

So this project solves this problem using 3 approaches.

• using js + tesseract.js + ollama (Qwen2.5) ( interacts with Tesseract ) inside v1/
• using shell + Tesseract ( c++ package which has to be installed in your system first ) inside v2
• using python + googletrans lib ( not-recommended ) inside v3/

Licenses

Installation

there are 2 ways/approaches to run this application, v1 and v2. Both of this application has their own set up procedure.

  git clone https://github.com/abhaysinghs772/pdf-ocr-parser.git

move to the cloned folder

  cd pdf-ocr-parser/

v1 [ using js and ollama (Qwen2.5) model ]

  cd v1/

v2 [ using raw shell scripts and Tesseract (c++ package) ]

  cd v3/

v3 [ using google-trans python lib ( NOT-RECOMMENDED )]

  cd v2/

Acknowledgements

• tesseract
• ollama
• QwenLm
• Kannada-OCR-test project by really cool IISc guys

Feedback

If you have any feedback or any issue, then please feel free to open the issues, or reach out to me directly at abhaysinghs772@gmail.com