IFormUnderstanding<T>
Interfaces · AiDotNet.Document.Interfaces
Interface for form understanding models that extract structured fields from documents.
Implements: IDocumentModel<T>
Known implementations: PICK<T>, TRIE<T>
For Beginners
Form understanding is like having someone read a form and fill out a digital version. The model finds:
- Field labels and their values (e.g., “Name: John Smith”)
- Checkboxes and whether they’re checked
- Signatures and their locations
Example usage:
How It Works
Form understanding models extract key-value pairs, checkboxes, signatures, and other structured information from forms, invoices, receipts, and similar documents.
Methods
| Method | Summary |
|---|---|
Dictionary<string, string> ExtractKeyValuePairs(Tensor<T> documentImage) |
Extracts key-value pairs from a document. |
FormFieldResult<T> ExtractFormFields(Tensor<T> documentImage) |
Extracts form fields from a document image. |
FormFieldResult<T> ExtractFormFields(Tensor<T> documentImage, double confidenceThreshold) |
Extracts form fields with a custom confidence threshold. |
IEnumerable<CheckboxResult<T>> DetectCheckboxes(Tensor<T> documentImage) |
Detects checkboxes and their states in a document. |
IEnumerable<SignatureResult<T>> DetectSignatures(Tensor<T> documentImage) |
Detects signatures in a document. |