This invention describes a system for automatically extracting specific pieces of information from electronic documents, particularly responses to questions on assessments. It works by letting a user define a template on an image of a document, marking where a specific piece of information is located. Once this template is set up, the system can then automatically find and pull out that same information from other similar documents, making the process of data extraction more efficient.
Why it matters: Filed before the widespread adoption of advanced AI for document understanding. Modern computer vision and large language models can now automate or significantly simplify the process of defining data segment positions and extracting information from varied document layouts, potentially making the template generation and extraction more robust and less manual than in 2018.
AI gives you a few directions you could take this. Pick one, and we check whether your version is different enough to patent, then write the filing.
Reinvent this with AI