Optical Character Recognition (OCR) technology bridges the physical and digital worlds by converting text within images into machine-readable data. For .NET developers, the ability to extract text from images in C# is essential for building intelligent document processing, automated data entry, and accessibility solutions.

In this article, we’ll explore how to implement OCR in C# using the Spire.OCR for .NET library, covering basic extraction, advanced features like coordinate tracking, and best practices to ensure accuracy and efficiency.

Table of Contents:

• Understanding OCR and Spire.OCR
• Setting Up Your OCR Environment
• Basic Recognition: Extract Text from Images in C#
• Advanced Extraction: Extract Text with Coordinates in C#
• Tips to Optimize OCR Accuracy
• FAQs (Supported Languages and Image Formats)
• Conclusion & Free License

Understanding OCR and Spire.OCR

What is OCR?

OCR technology analyzes images of text - such as scanned documents, screenshots, or photos - and converts them into text strings that can be edited, searched, or processed programmatically.

Why Spire.OCR Stands Out?

Spire.OCR for .NET is a powerful, developer-friendly library that enables highly accurate text recognition from images in C# applications. Key features include:

• Support for multiple languages (English, Chinese, Japanese, etc.).
• High accuracy recognition algorithms optimized for various fonts and styles.
• Text coordinate extraction for precise positioning.
• Batch processing capabilities.
• Compatibility with .NET Framework and .NET Core.

Setting Up Your OCR Environment

Before diving into the C# code for image to text OCR operations, configure your development environment first:

1. Install via NuGet:

Open the NuGet Package Manager in Visual Studio. Search for "Spire.OCR" and install the latest version in your project. Alternatively, use the Package Manager Console:

Install-Package Spire.OCR

2. Download OCR Models:

Spire.OCR relies on pre-trained models to recognize image text. Download the model files for your operating system:

• Windows x64
• Linux x64
• MacOS 10.15 and later

After downloading, extract to a directory (e.g., F:\OCR Model\win-x64)

Important Note: Remember to change the platform target of your solution to x64 as Spire.OCR only supports 64-bit platforms.

Basic Recognition: Extract Text from Images in C#

Let’s start with a simple example that demonstrates how to read text from an image using Spire.OCR.

C# code to get text from an image:

using Spire.OCR;
using System.IO;

namespace OCRTextFromImage
{
    internal class Program
    {
        static void Main(string[] args)
        {

            // Create an instance of the OcrScanner class
            OcrScanner scanner = new OcrScanner();

            // Create an instance of the ConfigureOptions class
            ConfigureOptions configureOptions = new ConfigureOptions();

            // Set the path to the OCR model
            configureOptions.ModelPath = "F:\\OCR Model\\win-x64";

            // Set the language for text recognition. (The default is English.)
            configureOptions.Language = "English";

            // Apply the configuration options to the scanner
            scanner.ConfigureDependencies(configureOptions);

            // Scan image and extract text
            scanner.Scan("sample.png");

            // Save the extracted text to a txt file
            string text = scanner.Text.ToString();
            File.WriteAllText("output.txt", text);
        }
    }
}

Code Explanation:

• OcrScanner: Core class for text recognition.
• ConfigureOptions: Sets OCR parameters:
  • ModelPath: Specifies the path to the OCR model files.
  • Language: Defines the recognition language (e.g., "English", "Chinese").
• Scan(): Processes image and extracts text using the configured settings.

Output:

This C# code processes an image file (sample.png) and saves the extracted text to a text file (output.txt) using File.WriteAllText().

Advanced Extraction: Extract Text with Coordinates in C#

In many cases, knowing the position of extracted text within an image is as important as the text itself - for example, when processing invoices, forms, or structured documents. Spire.OCR allows you to extract not just text but also the coordinates of the text blocks, enabling precise analysis.

C# code to extract text with coordinates from an Image:

using Spire.OCR;
using System.Collections.Generic;
using System.IO;

namespace OCRWithCoordinates
{
    internal class Program
    {
        static void Main(string[] args)
        {

            // Create an instance of the OcrScanner class
            OcrScanner scanner = new OcrScanner();

            // Create an instance of the ConfigureOptions class
            ConfigureOptions configureOptions = new ConfigureOptions();

            // Set the path to the OCR model
            configureOptions.ModelPath = "F:\\OCR Model\\win-x64";

            // Set the language for text recognition. (The default is English.)
            configureOptions.Language = "English";

            // Apply the configuration options to the scanner
            scanner.ConfigureDependencies(configureOptions);

            // Extract text from an image
            scanner.Scan("invoice.png");

            // Get the OCR result text
            IOCRText text = scanner.Text;

            // Create a list to store information
            List<string> results = new List<string>();

            // Iterate through each block of the OCR result text
            foreach (IOCRTextBlock block in text.Blocks)
            {
                // Add the text of each block and its location information to the list
                results.Add($"Block Text: {block.Text}");
                results.Add($"Coordinates: {block.Box}");
                results.Add("---------");
            }

            // Save the extracted text with coordinates to a txt file
            File.WriteAllLines("ExtractWithCoordinates.txt", results);
        }
    }
}

Critical Details

• IOCRText: Represents the entire OCR result.
• IOCRTextBlock: Represents a block of contiguous text (e.g., a paragraph, line, or word).
• IOCRTextBlock.Box: Contains the rectangular coordinates of the text block:
  • X (horizontal position)
  • Y (vertical position)
  • Width
  • Height

Output:

This C# code performs OCR on an image file (invoice.png), extracting both the recognized text and its position coordinates in the image, then saves this information to a text file (ExtractWithCoordinates.txt).

Tips to Optimize OCR Accuracy

To ensure reliable results when using C# to recognize text from images, consider these best practices:

• Use high-resolution images (300 DPI or higher).
• Preprocess images (e.g., resize, deskew) for better results.
• Ensure correct language settings correspond to the text in image.
• Store OCR models in a secure, accessible location.

FAQs (Supported Languages and Image Formats)

Q1: What image formats does Spire.OCR support?

A: Spire.OCR supports all common formats:

• PNG
• JPEG/JPG
• BMP
• TIFF
• GIF

Q2: What languages does Spire.OCR support?

A: Multiple languages are supported:

• English (default)
• Chinese (Simplified and Traditional)
• Japanese
• Korean
• German
• French

Q3: Can I use Spire.OCR in ASP.NET Core applications?

A: Yes. Supported environments:

• .NET Framework 2.0+
• .NET Standard 2.0+
• .NET Core 2.0+
• .NET 5

Q4: Can Spire.OCR extract text from scanned PDFs in C#?

A: The task requires the Spire.PDF integration to convert PDFs to images or extract images from scanned PDFs first, and then use the above C# examples to get text from the images.

Conclusion & Free License

Spire.OCR for .NET provides a powerful yet straightforward solution for extracting text from images in C# applications. Whether you’re building a simple tool to convert images to text or a complex system for processing thousands of invoices, by following the techniques and best practices outlined in this guide, you can integrate OCR functionality into your C# applications with ease.

Request a 30-day trial license here to get unlimited OCR capabilities and unlock valuable information trapped in visual format.