Template extraction

Sync Extraction

/ocrsvc/v{ver}/TemplateExtraction is called when user want to from the specified file formats like .jpg or .jpeg or .png.

The table shows the required values:

Component	Description
API Method	/ocrsvc/v{ver}/TemplateExtraction
Input	ver - The version of the IDP ocrDocument - Processes .jpg, .jpeg, and .png files as input. documentType - Type of the Document templateFile - The file of JSON type which represents the template of the document
Output	The response is of JSON type

Input Template

Template file is used to retrieve entities, tables, and key-value pairs using different techniques, such as regular expressions and region-based approaches. Below are some commonly used extractor types:

• Regular Expression Extractor (REExtractor): Extracts specific entities from text using predefined regular expression patterns. This is useful for structured data formats like dates, phone numbers, and invoice numbers.
• Region of Interest Extractor (ROIExtractor): Identifies and extracts entities based on predefined regions within a document. This method is useful when the position of the data is consistent across documents.
• Table Extraction using Regular Expressions (TableREExtractor): Extracts tabular data using pattern-based rules. This approach is effective for structured tables where patterns can be defined to locate rows and columns.
• Table Extraction using Region of Interest (TableROIExtractor): Retrieves tabular data by defining specific regions in a document. This is beneficial for documents where tables have a fixed layout.
• Key-Value Pair Extraction in Tables (KeyValueTableREExtractor): Extracts key-value pairs in a tabular format using regular expressions. This is particularly useful for structured forms where key-value relationships are consistently formatted.

Response output

{
  "Fields": [
    {
      "FieldName": "Name of the extracted entity", 
      "OCR_text": "Extracted text from OCR", 
      "Confidence": "Confidence score of extraction", 
      "PageNo": "Page number where the entity was found"
    }
  ],
  "Tables": [
    {
      "TableRows": [
        [
          {
            "ColumnName": "Column name of the table", 
            "OCR_text": "Extracted text from OCR", 
            "Geometry": [Left, Top, Width, Height] 
          }
        ]
      ],
      "PageNo": "Page number where the table was found"
    }
  ],
  "_metadata": {
    "TotalFields": "Total number of extracted fields", 
    "TotalTables": "Total number of extracted tables", 
"Confidence": "Overall confidence score",
"TaskID": "Unique identifier for the OCR processing job",
"OcrProvider": "Name of the OCR service provider used for extraction",
"TenantID": "Identifier for the tenant or client using the service",
    "NumberOfPages": "Total number of pages in the document"
  }
}

Async Extraction

/ocrsvc/v{ver}/AysncTemplateExtraction is called when called when user want to from the specified file formats like .pdf .

The table shows the required values:

Component	Description
API Method	/ocrsvc/v{ver}/AysncTemplateExtraction
Input	ver - The version of the IDP ocrDocument - Processes .pdf files as input. documentType - Type of the Document templateFile – The file of JSON type which represents the template of the document
Output	Task ID is generated.

Using the Task ID, user need to submit the Task ID in /ocrsvc/v{ver}/GetJobResult. This API returns the job result for all the Async API for the given TaskID

Response output

{
  "Fields": [
    {
      "FieldName": "Name of the extracted entity", 
      "OCR_text": "Extracted text from OCR", 
      "Confidence": "Confidence score of extraction", 
      "PageNo": "Page number where the entity was found"
    }
  ],
  "Tables": [
    {
      "TableRows": [
        [
          {
            "ColumnName": "Column name of the table", 
            "OCR_text": "Extracted text from OCR", 
            "Geometry": [Left, Top, Width, Height] 
          }
        ]
      ],
      "PageNo": "Page number where the table was found"
    }
  ],
  "_metadata": {
    "TotalFields": "Total number of extracted fields", 
    "TotalTables": "Total number of extracted tables", 
"Confidence": "Overall confidence score",
"TaskID": "Unique identifier for the OCR processing job",
"OcrProvider": "Name of the OCR service provider used for extraction",
"TenantID": "Identifier for the tenant or client using the service",
    "NumberOfPages": "Total number of pages in the document"
  }
}