Faster and More Reliable Info Extraction from Documents
Manual information extraction from documents is tedious and prone to error. For documents containing fine-grained table data (e.g. delivery notes or invoices), traditional information extraction software falls short and needs manual rework. ETH-spinoff BLP Digital AG combines deep learning algorithms for vision and NLP to fuse visual features of documents with its textual information. This solution is very accurate and offers high cost savings.