Skip to content

Form Field Extraction Process

Form field extraction is a key process in DocuSnap-Frontend that involves capturing form images, processing them, and extracting field information. This process is similar to document scanning but has specific differences in backend processing and data modeling.

Detailed Form Processing Flow

1. Form Capture Phase

  • User captures form images through the CameraCaptureScreen
  • Image data is passed to the ImageProcessingScreen
  • The camera interface provides guidance for optimal form capture
  • Multiple images can be captured for multi-page forms

2. Image Processing Phase

  • Similar to document scanning, images undergo edge detection, perspective correction, and enhancement
  • Processing is optimized specifically for form field recognition
  • Enhanced images are passed to the FormDetailScreen

3. Form Creation Phase

  • FormDetailScreen calls DocumentViewModel (shared ViewModel) to create a form object
  • User inputs form name, description, and tags
  • ViewModel calls Repository to save the form
  • Form metadata is associated with the processed images

4. Field Extraction Request Phase

  • Repository calls BackendApiService to send the form to the backend for field extraction
  • Backend returns a job ID, and a job record is created
  • JobPollingService begins polling for job status
  • The form is marked as "processing" in the local database

5. Field Extraction Result Processing Phase

  • When the job completes, JobPollingService retrieves and decrypts the results
  • Results include recognized form fields (names and values)
  • Repository updates the form entity, adding the extracted fields
  • The form status is updated to reflect the extraction completion

6. Field Editing Phase

  • FormDetailScreen displays the extracted fields
  • User can edit, confirm, or add fields
  • Changes are saved to the database through the ViewModel
  • Field validation rules can be applied during editing

7. Form Completion Phase

  • After editing, the form is marked as processed
  • Form data can be exported or integrated with other systems
  • Form usage statistics are updated
  • Form is available for future reference and searching

Differences from Document Scanning

While the form field extraction process shares similarities with document scanning, there are key differences:

  1. Field Recognition Focus:
  2. Forms focus on structured field extraction
  3. Documents focus on general text extraction

  4. Form processing identifies field names and values as pairs

  5. User Interaction:

  6. Form processing typically requires more user verification and editing
  7. Field values often need validation against specific rules

  8. Form fields have relationships that need to be preserved

  9. Backend Processing:

  10. Form processing uses specialized field extraction algorithms
  11. Form templates can be applied to improve recognition

  12. Field extraction includes confidence scores for user verification

  13. Data Model:

  14. Forms have a structured formFields collection
  15. Each field has a name, value, and extraction status
  16. Forms track field relationships and dependencies

Integration with Other Systems

The form field extraction process is designed to integrate with other systems:

  1. Data Export:
  2. Extracted form data can be exported in various formats (JSON, CSV, etc.)

  3. Export includes both field data and form metadata

  4. Template Learning:

  5. Frequently processed form types can be learned as templates

  6. Templates improve future extraction accuracy for similar forms

  7. Workflow Integration:

  8. Extracted form data can trigger workflows in other systems
  9. Integration APIs allow for seamless data transfer

This form field extraction process demonstrates how DocuSnap-Frontend combines automated processing with user interaction to provide an efficient and accurate form data capture and management solution.