01 / OVERVIEW
Overview
This repository is a fork of sdcb/SimdPaddleOCR, a pure C# PP-OCRv6 inference library. Upstream uses a managed ONNX interpreter and SIMD kernels for detection, orientation classification, and recognition. Its core API accepts pixel memory directly rather than requiring a particular image decoder. Review upstream documentation and licensing when evaluating it.
02 / CAPABILITIES
Key capabilities
Text detection, orientation classification, and recognition
A managed ONNX interpreter and SIMD kernels
Pixel-memory input without a mandatory image-decoding library
03 / POSSIBILITIES
Ideas to explore
- Explore image-text extraction and OCR workflows in .NET
- Compare integrations with different pixel layouts and image decoders
- Study upstream code and contribute documentation, tests, or inference improvements
04 / GET STARTED
Getting started
Before you begin
- Prepare .NET 10, the core package, and models following upstream documentation
- Provide pixel memory in a supported layout
- Review upstream and model licensing terms
Get the project
git clone https://github.com/hnlyf168/SimdPaddleOCR.gitNext steps
- Start with the upstream README for packages, models, and input layouts.
- Choose an image decoder and prepare pixels in a supported format.
- Run the upstream OCR example, then compare this fork with upstream if researching changes.
05 / SCOPE & STATUS
Current boundaries
This is a fork of sdcb/SimdPaddleOCR. Original authorship and licensing belong to the upstream project; it is not presented as an original project here.
Attribution
This repository is a fork. See the upstream repository for original authorship and licensing.
sdcb/SimdPaddleOCR ↗This introduction is based on the public README. Consult the repository for current versions, licenses, and supported features.
Updated 2026-10-27