High-Accuracy Data Extraction from OCR

I need a reliable solution to interpret specific tire data from OCR results. The data of interest includes:
- Country
- Size
- Model
- DOT code
- Speed index

Current OCR results are in JPG format. Automated software has been used but suffers from:
- Low accuracy
- Incomplete data extraction

I’m aiming for a high accuracy rate (99% is desirable).

Ideal Skills and Experience:
- Expertise in OCR and data extraction
- Proficiency in handling JPG input
- Experience with high-accuracy automation solutions
- Ability to troubleshoot and improve existing automated software