Document Intelligence read model

Document Intelligence Read Optical Character Recognition (OCR) model runs at a higher resolution than Azure AI Vision Read and extracts print and handwritten text from PDF documents and scanned images. It also includes support for extracting text from Microsoft Word, Excel, PowerPoint, and HTML documents. It detects paragraphs, text lines, words, locations, and languages. The Read model is the underlying OCR engine for other Document Intelligence prebuilt models like Layout, General Document, Invoice, Receipt, Identity (ID) document, Health insurance card, W2 in addition to custom models.

What is OCR for documents?

Optical Character Recognition (OCR) for documents is optimized for large text-heavy documents in multiple file formats and global languages. It includes features like higher-resolution scanning of document images for better handling of smaller and dense text; paragraph detection; and fillable form management. OCR capabilities also include advanced scenarios like single character boxes and accurate extraction of key fields commonly found in invoices, receipts, and other prebuilt scenarios.

Development options

Input requirements

• Supported file formats:

  Model	PDF	Image: JPEG/JPG, PNG, BMP, TIFF, HEIF	Microsoft Office: Word (DOCX), Excel (XLSX), PowerPoint (PPTX), HTML
  Read	✔	✔	✔
  Layout	✔	✔	✔ (2024-07-31-preview, 2024-02-29-preview, 2023-10-31-preview)
  General Document	✔	✔
  Prebuilt	✔	✔
  Custom extraction	✔	✔
  Custom classification	✔	✔	✔ (2024-07-31-preview, 2024-02-29-preview)

• For best results, provide one clear photo or high-quality scan per document.

• For PDF and TIFF, up to 2,000 pages can be processed (with a free tier subscription, only the first two pages are processed).

• The file size for analyzing documents is 500 MB for paid (S0) tier and 4 MB for free (F0) tier.

• Image dimensions must be between 50 pixels x 50 pixels and 10,000 pixels x 10,000 pixels.

• If your PDFs are password-locked, you must remove the lock before submission.

• The minimum height of the text to be extracted is 12 pixels for a 1024 x 768 pixel image. This dimension corresponds to about 8 point text at 150 dots per inch (DPI).

• For custom model training, the maximum number of pages for training data is 500 for the custom template model and 50,000 for the custom neural model.

  • For custom extraction model training, the total size of training data is 50 MB for template model and 1 GB for the neural model.

  • For custom classification model training, the total size of training data is 1 GB with a maximum of 10,000 pages. For 2024-07-31-preview and later, the total size of training data is 2 GB with a maximum of 10,000 pages.

Get started with Read model

Try extracting text from forms and documents using the Document Intelligence Studio. You need the following assets:

• An Azure subscription—you can create one for trial

• A Document Intelligence instance in the Azure portal. You can use the free pricing tier (F0) to try the service. After your resource deploys, select Go to resource to get your key and endpoint.

Sample document processed with Document Intelligence Studio

1. On the Document Intelligence Studio home page, select Read.

2. You can analyze the sample document or upload your own files.

3. Select the Run analysis button and, if necessary, configure the Analyze options:

   

   Try Document Intelligence Studio.

Supported languages and locales

See our Language Support—document analysis models page for a complete list of supported languages.

Data extraction

Searchable PDF

The searchable PDF capability enables you to convert an analog PDF, such as scanned-image PDF files, to a PDF with embedded text. The embedded text enables deep text search within the PDF's extracted content by overlaying the detected text entities on top of the image files.

Use searchable PDF

To use searchable PDF, make a POST request using the Analyze operation and specify the output format as pdf:

POST /documentModels/prebuilt-read:analyze?output=pdf
{...}
202

Poll for completion of the Analyze operation. Once the operation is complete, issue a GET request to retrieve the PDF format of the Analyze operation results.

Upon successful completion, the PDF can be retrieved and downloaded as application/pdf. This operation allows direct downloading of the embedded text form of PDF instead of Base64-encoded JSON.

// Monitor the operation until completion.
GET /documentModels/prebuilt-read/analyzeResults/{resultId}
200
{...}

// Upon successful completion, retrieve the PDF as application/pdf.
GET /documentModels/prebuilt-read/analyzeResults/{resultId}/pdf
200 OK
Content-Type: application/pdf

Pages

The pages collection is a list of pages within the document. Each page is represented sequentially within the document and includes the orientation angle indicating if the page is rotated and the width and height (dimensions in pixels). The page units in the model output are computed as shown:

"pages": [
    {
        "pageNumber": 1,
        "angle": 0,
        "width": 915,
        "height": 1190,
        "unit": "pixel",
        "words": [],
        "lines": [],
        "spans": []
    }
]

# Analyze pages.
for page in result.pages:
    print(f"----Analyzing document from page #{page.page_number}----")
    print(
        f"Page has width: {page.width} and height: {page.height}, measured with unit: {page.unit}"
    )

"pages": [
    {
        "pageNumber": 1,
        "angle": 0,
        "width": 915,
        "height": 1190,
        "unit": "pixel",
        "words": [],
        "lines": [],
        "spans": []
    }
]

Select pages for text extraction

For large multi-page PDF documents, use the pages query parameter to indicate specific page numbers or page ranges for text extraction.

Paragraphs

The Read OCR model in Document Intelligence extracts all identified blocks of text in the paragraphs collection as a top level object under analyzeResults. Each entry in this collection represents a text block and includes the extracted text ascontent and the bounding polygon coordinates. The span information points to the text fragment within the top-level content property that contains the full text from the document.

"paragraphs": [
    {
        "spans": [],
        "boundingRegions": [],
        "content": "While healthcare is still in the early stages of its Al journey, we are seeing pharmaceutical and other life sciences organizations making major investments in Al and related technologies.\" TOM LAWRY | National Director for Al, Health and Life Sciences | Microsoft"
    }
]

Text, lines, and words

The Read OCR model extracts print and handwritten style text as lines and words. The model outputs bounding polygon coordinates and confidence for the extracted words. The styles collection includes any handwritten style for lines if detected along with the spans pointing to the associated text. This feature applies to supported handwritten languages.

For Microsoft Word, Excel, PowerPoint, and HTML, Document Intelligence Read model v3.1 and later versions extracts all embedded text as is. Texts are extrated as words and paragraphs. Embedded images aren't supported.

"words": [
    {
        "content": "While",
        "polygon": [],
        "confidence": 0.997,
        "span": {}
    },
],
"lines": [
    {
        "content": "While healthcare is still in the early stages of its Al journey, we",
        "polygon": [],
        "spans": [],
    }
]

# Analyze lines.
for line_idx, line in enumerate(page.lines):
    words = line.get_words()
    print(
        f"...Line # {line_idx} has {len(words)} words and text '{line.content}' within bounding polygon '{format_polygon(line.polygon)}'"
    )

    # Analyze words.
    for word in words:
        print(
            f"......Word '{word.content}' has a confidence of {word.confidence}"
        )

"words": [
    {
        "content": "While",
        "polygon": [],
        "confidence": 0.997,
        "span": {}
    },
],
"lines": [
    {
        "content": "While healthcare is still in the early stages of its Al journey, we",
        "polygon": [],
        "spans": [],
    }
]

# Analyze lines.
if page.lines:
    for line_idx, line in enumerate(page.lines):
        words = get_words(page, line)
        print(
            f"...Line # {line_idx} has {len(words)} words and text '{line.content}' within bounding polygon '{line.polygon}'"
        )

        # Analyze words.
        for word in words:
            print(f"......Word '{word.content}' has a confidence of {word.confidence}")

"words": [
    {
        "content": "While",
        "polygon": [],
        "confidence": 0.997,
        "span": {}
    },
],
"lines": [
    {
        "content": "While healthcare is still in the early stages of its Al journey, we",
        "polygon": [],
        "spans": [],
    }
]

Handwritten style for text lines

The response includes classifying whether each text line is of handwriting style or not, along with a confidence score. For more information, see handwritten language support. The following example shows an example JSON snippet.

"styles": [
{
    "confidence": 0.95,
    "spans": [
    {
        "offset": 509,
        "length": 24
    }
    "isHandwritten": true
    ]
}

If you enabled the font/style addon capability, you also get the font/style result as part of the styles object.

Next steps

Complete a Document Intelligence quickstart:

Explore our REST API: