The ‘AI-based extraction of invoice items’ (ExRePo) project aims to use artificial intelligence (AI) to automatically extract information from business documents, in particular invoices in the context of agriculture. The focus is particularly on invoice item data, i.e. information such as product descriptions, quantities and prices. These are to be provided in a machine-readable form in order to make business processes, e.g. in accounting and auditing, more efficient and error-free. The project uses machine learning to develop models that enable the automated processing of various invoice formats. An innovative aspect of the project is the generation of synthetic invoices to improve the training of AI models. In addition, the extracted data will be used to develop innovative services in an agricultural context. Examples of this include yield forecasts and recommendations for seeds and fertilisers based on the information obtained. These additional applications can support agricultural businesses in optimising their decisions regarding cultivation and resource management, which offers both ecological and economic benefits.
The project is being realised by the Leuphana University of Lüneburg in cooperation with LAND-DATA GmbH and is funded as part of the ERDF guideline "Innovations by universities and research institutions".
Prof Dr Burkhardt Funk will be happy to answer any questions you may have.