Abstract
This master's thesis examines the use of artificial intelligence (AI) and machine learning (ML) for the detection and processing of unstructured incoming invoices. Traditional methods of invoice processing are often labor-intensive and prone to errors,especially when dealing with unstructured formats such as scanned invoices in PDF
format. The thesis describes the use of Optical Character Recognition (OCR) techniques combined with Natural Language Processing (NLP) methods, as well as approaches related to object detection, to extract relevant information from these
documents. A prototype system was developed using cloud technologies and evaluated with a dataset of real invoices. The mapping of the identified invoice fields to the
SAP master data is implemented using lazy learning technologies, ensuring that the
system achieves an improved recognition rate for future invoices.
| Date of Award | 2024 |
|---|---|
| Original language | German (Austria) |
| Supervisor | Stefan Hinterholzer (Supervisor) |