Aspose.Ocr.Cpp 20.7.1

Optical Character Recognition (OCR) C++ API

Aspose.OCR for C++ is a standalone OCR API that enhances your C++ apps to perform the OCR operation on JPEG, PNG, & BMP images for extraction of English content.

Image OCR API Features

• Recognize characters from images.
• Calculate the skew angle of images.
• Currently it supports English language.
• Auto-align image and autodetect text areas.
• Get a single line and recognize it.

New Features & Enhancements in Version 20.7

• New ML model with improved recognition.
• Ability to recognize text in the tables.
• Added image recognition using GPU and CUDA technology (available in separate assembly Aspose.Ocr.Cpp.Gpu).

For the detailed notes, pelase visit Aspose.OCR for C++ 20.7 Release Notes.

Read Image Formats for OCR

Raster Formats: JPEG, BMP ( 32bit and 16bit images are not supported), PNG

Supported Charset for Recognition

EN

Pre-requisites

Aspose OCR library requires onnxruntime.dll in the system path.

Getting Started with Aspose.OCR for C++

Are you ready to give Aspose.OCR for C++ a try? Simply execute Install-Package Aspose.OCR.Cpp from Package Manager Console in Visual Studio to fetch the NuGet package. If you already have Aspose.OCR for C++ and want to upgrade the version, please execute Update-Package Aspose.OCR.Cpp to get the latest version.

Execute the following code snippet to see how Aspose.OCR API performs with your own samples or check the GitHub Repository for other common usage scenarios.

Perform OCR on PNG Image via C++ Code

// For complete examples and data files, please go to https://github.com/aspose-ocr/Aspose.OCR-for-C
std::string image_path = "../Data/Source/sample.png";
const wchar_t* result = aspose::ocr::recognize_page(image_path.c_str());
std::wcout << result << L'\n';

Evaluation Version Limitation

The evaluation version of Aspose.OCR for C++ limits the number of characters extracted from an image to 300.

Home | Product Page | Docs | Demos | API Reference | Examples | Blog | Free Support | Temporary License