Technology Sharing

한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina

I. Introduction

Since the project requires OCR recognition capabilities and needs to support private deployment, this article will compare some open source OCR recognition tools on the market, select the OCR that suits the project needs, and further study/train the corresponding OCR model in the future.
The main OCR recognitions are: Tesseract_OCR, PaddleOCR, EasyOCR, dddd_ocr, CnOCR
Note: The following image tests are used as follows

2. Tesseract_OCR

Pillow is a free and open source image processing library that can be used to read, manipulate and save a variety of image files. Tesseract-OCR is a powerful optical character recognition engine that can recognize text in images offline and accurately. It should be used in conjunction with the locally installed tesseract-ocr.exe file.
Tesseract-OCR Features:

• Tesseract supports UTF-8 encoding format and can recognize more than 100 languages ​​​​out of the box
• Tesseract supports multiple output formats: plain text, hOCR (HTML), PDF, etc.
• The official recommendation is to provide high-quality images for better OCR results.
• Tesseract is trained to recognize other languages. For specific training methods, please refer to the official documentation: https://tesseract-ocr.github.io/tessdoc/

2.1. Installation process

Installation Environment：