Extracting Table Data from Scanned Images Using IronOCR : Live Demo Recap

Extracting data from scanned images is a common challenge, especially when it involves structured data like tables. With IronOCR's advanced machine learning capabilities, you can now seamlessly extract table data including cell values and their positions. In this demo, Shadman Majid, Software Sales Engineer, walks through the code implementation step-by-step, while Anne Lazarakis, Sales and Marketing Director, shares real-world use cases from Iron Software customers.

Real-World Use Cases

Explained by Anne Lazarakis, Sales and Marketing Director*

Insurance Claim Processing (Opyn Market)

In the highly regulated healthcare insurance industry in the U.S., companies like Opyn Market still receive many documents via fax. These scanned documents often contain tabular data that must be accurately extracted and entered into internal systems. With IronOCR, they’re able to automate this process, reducing manual work and eliminating the potential for human error.

Logistics & Food Distribution (iPAP)

iPAP, the largest cheese distributor in the U.S., uses IronOCR to manage over 200 client orders. Their invoices come in various formats with inconsistent table layouts. IronOCR helps them extract purchase order numbers, shipment dates, and item details from scanned documents efficiently, even with varied formatting. This automation has saved them between $40,000 and $45,000 annually.

Technical Overview

Live Coding Session With Shadman Majid, Software Sales Engineer*

IronOCR uses proprietary machine learning models to detect and extract table data from scanned documents. This feature supports:

• Extraction of table cells and coordinates
• OCR of scanned images and multi-frame PDFs
• Compatibility with C#, VB.NET, .NET Standard, .NET Framework, and .NET Core

To access this functionality, you'll need:

• IronOCR NuGet package
• IronOcr.Extensions.AdvancedScanning NuGet package for table detection via ML models

These packages include the trained ML models necessary for table structure detection and accurate OCR.

Example Code for Extracting Tables

Below is a sample C# code snippet that demonstrates how to use IronOCR for extracting table data from images:

// Import the necessary IronOCR namespaces
using IronOcr;

// Initialize the IronTesseract to handle OCR processes
var Ocr = new IronTesseract();

// Load the image containing the table
using (var input = new OcrInput("invoice.jpg"))
{
    // Perform OCR and extract text data including tables
    var result = Ocr.Read(input);

    // Iterate through each page in the document
    foreach (var page in result.Pages)
    {
        // Iterate through each table found on the page
        foreach (var table in page.Tables)
        {
            Console.WriteLine("Table found:");
            // Iterate through each row in the table
            foreach (var row in table.Rows)
            {
                // Convert the row of cells to a comma-separated string
                var cells = string.Join(", ", row.Cells.Select(cell => cell.Text));
                Console.WriteLine(cells);
            }
        }
    }
}

// Import the necessary IronOCR namespaces
using IronOcr;

// Initialize the IronTesseract to handle OCR processes
var Ocr = new IronTesseract();

// Load the image containing the table
using (var input = new OcrInput("invoice.jpg"))
{
    // Perform OCR and extract text data including tables
    var result = Ocr.Read(input);

    // Iterate through each page in the document
    foreach (var page in result.Pages)
    {
        // Iterate through each table found on the page
        foreach (var table in page.Tables)
        {
            Console.WriteLine("Table found:");
            // Iterate through each row in the table
            foreach (var row in table.Rows)
            {
                // Convert the row of cells to a comma-separated string
                var cells = string.Join(", ", row.Cells.Select(cell => cell.Text));
                Console.WriteLine(cells);
            }
        }
    }
}

' Import the necessary IronOCR namespaces
Imports IronOcr

' Initialize the IronTesseract to handle OCR processes
Private Ocr = New IronTesseract()

' Load the image containing the table
Using input = New OcrInput("invoice.jpg")
	' Perform OCR and extract text data including tables
	Dim result = Ocr.Read(input)

	' Iterate through each page in the document
	For Each page In result.Pages
		' Iterate through each table found on the page
		For Each table In page.Tables
			Console.WriteLine("Table found:")
			' Iterate through each row in the table
			For Each row In table.Rows
				' Convert the row of cells to a comma-separated string
				Dim cells = String.Join(", ", row.Cells.Select(Function(cell) cell.Text))
				Console.WriteLine(cells)
			Next row
		Next table
	Next page
End Using

• Loading an Image: The script begins by initializing the IronTesseract engine and loading an image file named invoice.jpg that you want to process.
• OCR Execution: It performs OCR on the input to extract text data, particularly focusing on any tables.
• Table Extraction: The script iterates through each detected table and its rows, outputting cell contents in a structured way.

Ensure you have installed the necessary NuGet packages for IronOCR before running this script.

Conclusion

IronOCR makes it easy to automate the extraction of complex table data from scanned documents. Whether you're in healthcare, logistics, finance, or manufacturing, this solution offers reliability, accuracy, and cost-saving efficiency. With just a few lines of code, you can eliminate manual data entry and reduce human error.

Want to see it in action? Book a live Demo with one of our engineers here.