👁 Image
Aspose.OCR-GPU 26.3.0

Optical character recognition (OCR) API for .NET

👁 Version 26.3.0
👁 Downloads

👁 banner

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

Integrate optical character recognition (OCR) into your .NET applications, websites, cloud services, or serverless functions written in C#, F# and VB.NET. Transform scanned documents into searchable PDFs, extract text from photos, illustrations, and screenshots, compare text across images, and many more.

Cross-platform

Aspose OCR library works everywhere and is suitable for any task:

• x64 and x86 architectures
• Windows, macOS, Linux, AWS, Azure, Google Cloud, Docker.
• .NET 9, .NET 8, .NET 7, .NET 6, .NET 5, .NET Core, .NET Standard, and .NET Framework.
• Desktop applications, console utilities, services, web sites, microservices, serverless functions.

You would like

Aspose offers one of the most powerful OCR API on the market:

• Universal – extracts text from any image, whether it's a high-quality scan, casual street photo, screenshot, or handwritten note.
• Reliable - delivers exceptional recognition accuracy, outperforming even the most powerful open source libraries such as Tesseract. Easily handles rotated, distorted, blurry or noisy images.
• Developer-friendly - you will only need a few lines of code to perform any OCR task. Choose between optimized, fully automated routines for effortless OCR processing or take full control over recognition, including session management and memory handling.
• Modular – maintains a lean and efficient application by selectively incorporating advanced features from our comprehensive resource repository.
• Multi-language - recognizes text in over 140 languages and seamlessly handles mixed-language content, even when composed of 5–6 different languages.
• Offline - all OCR tasks are performed locally, with no reliance on cloud or third-party services.

Try it yourself

Don't just take our word for it - try it yourself! Install the Aspose.OCR NuGet package and see it in action. With just five lines of C# code, you can convert an image to text:

Aspose.OCR.AsposeOcr recognitionEngine = new Aspose.OCR.AsposeOcr();
Aspose.OCR.OcrInput source = new Aspose.OCR.OcrInput(Aspose.OCR.InputType.SingleImage);
source.Add("your_image.png");
Aspose.OCR.OcrOutput results = recognitionEngine.Recognize(source);
Console.WriteLine(results[0].RecognitionText);

You can download sample projects to explore the advanced capabilities of the Aspose OCR library or visit the product page for live code sample.

Recognition languages

Our OCR library is ideal for global applications and multilingual documents. With support for nearly , it ensures highly accurate text extraction across various scripts and alphabets. We excel at processing mixed-language documents, even those containing 3+ languages like Japanese, English, and Cyrillic.

• Extended Latin (English, Spanish, Portuguese, French, German and other European languages)
• Cyrillic (Russian, Ukrainian, Bulgarian, and more)
• Arabic, Persian and Urdu
• Chinese
• Hindi and other Devanagari-based scripts
• Korean
• Japanese
• South Asian languages (Tamil, Telugu, Vietnamese, and the like)

Supported file formats

Aspose OCR can handle nearly any file format from scanners or cameras, including:

• JPEG, PNG, BMP, and GIF
• Single-page and multi-page TIFF
• PDF
• DjVu

All the mentioned file types can be read directly from web links. Additionally, our OCR library enables bulk recognition of images from folders and archives, making large-scale document processing faster and more efficient.

The recognition results are returned in popular file and data exchange formats, enabling easy storage, database import, or real-time analysis:

• Plain text
• Searchable PDF
• Microsoft Word or Excel documents
• HTML
• RTF
• ePUB
• JSON and XML
• hOCR

Image analysis and preprocessing

The Aspose.OCR library features fully automated image preprocessing, enabling it to extract text even in cases where other OCR engines, like Tesseract, fail to produce any results.

• Automatically detects image orientation, handling severely rotated and even upside-down images.
• Detects text background, recognizing inverted images.
• Automatically straightens page curls and corrects lens distortions.
• Removes dirt, spots, scratches, glare, and other noise.
• Automatically detects multi-column text and tables, recognizing and preserving their layout.

Licensing

You can begin using the Aspose OCR library immediately after installation. In evaluation mode (without providing a license), you can recognize text in any supported language and save recognition results in all supported formats, with a few limitations:

• If the number of characters in a recognized image exceeds 300, only the first 300 characters will be recognized.
• If the number of characters is fewer than 300, only the first 60% of the characters will be recognized.

To unlock all features, request a temporary license, which removes all limitations for 30 days. This gives you the opportunity to build a fully functional OCR application and evaluate Aspose.OCR before making your final purchase decision.

Learn more

• Online documentation.
• Real-life use cases.
• Free support forum.
• New feature requests and urgent assistance to ensure the library meets your specific needs.
• Other programming languages.