Understanding the document view: in-depth guide

Introduction to the document view

The document view is the central interface in smartextract where you can review, validate, and refine extracted data from your documents.
This powerful workspace provides a comprehensive set of tools for ensuring the accuracy of your extracted information and customizing models to better suit your unique requirements.

Navigating the document view interface

The document view is divided into three primary sections:

• left panel: Document thumbnail navigation displaying page previews for multi-page documents, allowing you to quickly jump between pages

• center panel: Full document display with highlighted bounding boxes indicating recognized data fields

• right panel: Structured display of all extracted fields organized by logical categories

Reviewing and correcting extracted data

smartextract’s AI models are highly accurate, but verification is sometimes necessary. The system makes it easy to identify and correct any extraction errors.

Validation mehtods:

1. Visual verification: Bounding boxes in the document show exactly where each piece of data was extracted from

2. Field highlighting: Click on any extracted field in the right panel to see its corresponding location highlighted in the document

3. Navigation links: Click on field labels to jump directly to their location in the document

Making corrections:

If you notice an incorrect extraction, you can easily fix it:

1. Click on the field value in the right panel

2. Edit the text directly in the field

3. The system automatically saves your correction

Customizing pre-built models

While smartextract provides several pre-built models, you can modify them to better align with your specific document formats:

Accessing model customization:

1. Click the "Customize AI model" button in the top right corner

2. The customization interface will appear in the right panel

Creating field groups:

Field groups organize related information into logical sections:

1. Click "New field group"

2. Provide a descriptive name (e.g., "Certificate Information")

3. Click "Next" to begin adding fields

Adding different field types:

smartextract supports multiple field types to accommodate various data formats:

Text fields:

1. Click "Add field"

2. Enter the field name (e.g., "Register Number")

3. Select "Text" as the field type

4. Click "Confirm"

Text fields are ideal for extracting document numbers, names, addresses, and other textual information.

Multiple choice fields:

1. Click "Add field"

2. Enter the field name (e.g., "Ventilation Type")

3. Select "Multiple Choice" as the field type

4. Enter all possible options separated by commas (e.g., "Natural, Mechanical, Hybrid")

5. Click "Confirm"

Multiple choice fields are excellent for standardizing data where only certain values are valid, such as categories, statuses, or classifications.

Date fields:

1. Click "Add field"

2. Enter the field name (e.g., "Certification Date")

3. Select "Date" as the field type

4. Click "Confirm"

Date fields automatically format extracted dates in a consistent manner.

Quantity fields:

1. Click "Add Field"

2. Enter the field name (e.g., "Total Area")

3. Select "Quantity" as the field type

4. Click "Confirm"

Quantity fields ensure that extracted values are properly recognized as numbers for calculations and reporting.

Creating inferred fields:

You can create fields for information that isn't explicitly stated in the document but can be inferred:

1. Add a new field (e.g., "Signature Status")

2. Select "Multiple Choice" as the field type

3. Enter the possible values (e.g., "Signed, Unsigned")

4. Click "Confirm"

The AI model will analyze the document and make a determination based on visual cues and content.

Handling extraction challenges

Missing fields

If critical information isn't being extracted:

1. Click "Customize AI Model"

2. Add the missing field to the appropriate field group

3. Save the model and allow it to reprocess the document

4. Verify the field is now being correctly extracted

Best practices for document review

• Review strategically: Focus on high-value fields and those with lower confidence scores first

• Maintain consistent naming: Use clear, descriptive field names that match across similar documents

• Group related fields: Organize fields logically to improve usability and extraction accuracy

• Document model changes: Keep notes on customizations for team knowledge sharing

Conclusion

The Document View in smartextract offers powerful tools for reviewing, correcting, and customizing your data extraction process. By understanding how to effectively use these features, you can ensure the highest level of accuracy while tailoring the system to your specific document processing needs.

Whether you're using pre-built models or creating custom extraction templates, smartextract puts you in control of your data extraction workflow, enabling you to process documents with confidence and precision.