Skip to content

README.md

Latest commit

 

History

History
67 lines (44 loc) · 2.37 KB

README.md

File metadata and controls

67 lines (44 loc) · 2.37 KB

node-paddle-ocr

Offline, Multi-language and Faster OCR in Terminal.

English | 简体中文

usage-png

📖 Introduction

An OCR terminal tool based on the open-source offline components of PaddleOCR v2.6 C++.

⚙️ Installation

  1. Clone this repo
  2. Run cd node-paddle-ocr
  3. Run npm run build
  4. Run npm link

🚀 Usage

ocr <image path> [options]

Options:
  -t, --target   Detection language specified as ISO 639-1 code         [string]
  -c, --code     List supported language codes                         [boolean]
  -v, --version  Show version number                                   [boolean]
  -h, --help     Show help                                             [boolean]

Examples:
  $ ocr              Identify pictures in the clipboard
  $ ocr ./image.png  Identify the image.png file in the current directory

⚠️ Compatibility

  • The system supports Win10/11 x64.
  • It is not recommended to use Win7 as the recognition engine is highly likely to be unable to run.
  • The CPU must have the AVX instruction set. Common home-use CPUs generally meet this condition.
AVX Supported product series Not supported
Intel Core series, Xeon series Atom, Itanium, Celeron, Pentium
AMD Products after the Bulldozer architecture such as Ryzen, Athlon, FX, etc. Products before the K10 architecture

🔗 Related

  • PaddleOCR-json - An OCR offline image text recognition command-line Windows program that outputs results in the form of JSON strings for convenient invocation by other programs. It provides APIs in various languages. Compiled by PaddleOCR C++.
  • PaddleOCR-json-node-api - A node.js API based on PaddleOCR-json.

🤝 Contribution

Contributions via Pull Requests or Issues are welcome.

📄 License

This project is licensed under the MIT License. See the LICENSE file for details.