01 / OVERVIEW
Overview
SimdPaddleOCR provides managed OCR with a pixel-memory API, allowing callers to choose their image decoder and prepare the documented models. Public NuGet packages provide core functionality and model components.
02 / CAPABILITIES
Key capabilities
01
Text detection, orientation classification, and recognition
02
Managed ONNX interpretation and SIMD kernels
03
Pixel-memory input and optional GPU backends
03 / POSSIBILITIES
Ideas to explore
- Extract text from images and documents in .NET
- Study managed ONNX and SIMD inference
- Compare pixel-layout and image-decoder integration
04 / GET STARTED
Getting started
Before you begin
- Prepare .NET, models, and image decoding as documented
- Verify supported formats, models, and GPU backends
Get the project
dotnet add package Sdcb.SimdPaddleOCRNext steps
- Install the core package and documented model packages.
- Supply pixel memory following an example and validate your images.
05 / SCOPE & STATUS
Current boundaries
This is the upstream repository maintained by sdcb, listed separately from the hnlyf168 fork. Core APIs do not decode image files; GPU use requires compatible drivers.
This introduction is based on the public README. Consult the repository for current versions, licenses, and supported features.
Updated 2026-10-27