Content-independent structural fingerprints for business forms
Generates layout/structure fingerprints from OCR feature output, then matches incoming forms against previously processed ones at high speed. Because content like company names, counterparty names, or numerical values is never used in the match, the system is unaffected by spelling variation or sender changes. Built for OCR post-processing pipelines that want to auto-apply previously registered form templates to similar future forms.
* Detailed information will be published in stages. For inquiries, please contact us.
