Optical Character Recognition (OCR)
Optical Character Recognition (OCR) is a technology that converts images of text — scanned documents, photos, or PDFs — into machine-readable text data. In the context of document verification, OCR enables the automatic extraction of identity information from supporting documents.
OCR is the foundational first step in any automated document verification process. When a user submits an ID card, proof of address, or bank statement, OCR technology analyses the image pixel by pixel to identify and extract printed or handwritten characters. Modern algorithms combine traditional pattern recognition with deep neural networks to achieve accuracy rates above 99% on standardised documents.
In the context of KYC and regulatory compliance, OCR dramatically reduces document processing times. Where a human operator would spend several minutes manually entering data from a document, OCR performs this extraction in seconds. This speed is essential for delivering a seamless customer experience while meeting AML regulatory requirements.
Advanced OCR solutions, such as those integrated into CheckFile, go beyond extracting raw text. They structure the extracted data (name, date of birth, document number, address) and map it to specific zones on the document, enabling cross-referencing checks and forgery detection.
Regulations
Real-world examples
- 1.An online bank uses OCR to automatically extract the name, date of birth, and ID card number during remote account opening, reducing onboarding time from 15 minutes to 30 seconds.
- 2.An accounting firm scans proof of address documents for new clients, and OCR extracts the address, account holder name, and issue date to directly populate the KYC file.
- 3.A letting agency uses OCR to read tenants' payslips and automatically verify that declared amounts are consistent with the extracted data.