A UIPath-based Approach for Automated Invoice Data Capture:A Comparative Study of Google Cloud Vision, OmniPage, and Tesseract OCR

The extraction of data from invoices requires accuracy combined with efficiency to enhance business financial operations and reduce human errors in daily processes. Our study establishes an invoice data extraction automation system through UIPath and tests three main OCR engines Google Cloud Vision OCR, OmniPage OCR and Tesseract OCR on 122 invoices which use different realistic format styles. Google Cloud Vision OCR demonstrates complete accuracy by delivering 100% precision for extracting data from all studied fields that include Invoice Number, Invoice Date and Invoice Amount. The invoice fields Invoice Number and Invoice Date show identical extraction proficiency at 100% with Tesseract OCR but the Invoice Amount extraction achieves 90.16% accuracy. The Invoice Date information retrieval rate reaches only 61.48% in OmniPage OCR while the program shows both high accuracy in Invoice Number extraction (99.18%) and medium accuracy in Invoice Amount extraction (89.34%). Our study demonstrates the benefits and constraints of the featured OCR engines along with valuable information that helps companies implement successful invoice automation approaches.