joshsoftware / joshsoftware/computer_vision_pocs

Evaluate OCR Models and Finalize Best Fit

Open
#2 0 comments 0 reactions 0 assignees View on GitHub
Dominant language
No language data
Stars
1
Forks
0
PR merge metrics
No merged PRs in 30d

Description

# Evaluate OCR Models and Finalize Best Fit (Refer: Wiki Report)

## Description

We have tested multiple OCR models to evaluate their accuracy, ease of use, and performance across English and Marathi text inputs.

The detailed results and observations are documented in the project wiki:

🔗 **[OCR Model Comparison Report](https://github.com/joshsoftware/computer_vision_pocs/wiki/OCR-Model-Report)**

---

## Summary of Findings

- **Tesseract**
Best overall performance. Accurate, lightweight, and easy to integrate.

- **EasyOCR**
Moderate performance; issues with structure and sentence clarity.

- **I2L-NOPOOL** & **DTrOCR**
Requires additional training or lacks proper documentation.

- **GPT-4o**
Requires GPT Plus subscription (not free).

---

## Additional Notes

- This issue is linked to the [OCR research wiki](https://github.com/joshsoftware/computer_vision_pocs/wiki/OCR-Model-Report) for ongoing reference.
- We may consider post-processing like word segmentation or spell correction in the future for further improvements.

Contributor guide

No contributing guide indexed for this repository

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.