Optical Character Recognition (OCR) is a important task in information.

However, the OCR perfromance can vary depending on image quality, font styles, or the language ofthe text.

Currently, many open-source OCR engine each of them give better performance

under a point-wise conditions. This internship aims to develop approaches for merging

the outputs of multiple OCR systems to improve the quality of the results.

The work will focus on:

• A comparative analysis of the performance of various OCR systems.

• Designing and implement a method for fusion (e.g., based on majority voting, probabilistic

models, or machine learning techniques).

• Evaluating the performance of the fused system on diverse datasets.