Page Segmentation: The easy and the hard way
An OCR scan of a whole page of a complex layout can be done two ways. The easy expensive one using an LLM or the more sophisticated one, which is harder to develop but…
These are side projects across the fields of machine learning and software design.
Some are quick espresso experiments, others are slow-brewed prototypes, and a few are already close to an MVP.
More prototypes and experiments can be found on this page.