I'm looking to use an LLM with unstructured data in an image that has hand written notes. Is it possible to extract the notes from the image and feed it into the LLM as text to then generate structured data? It feels similar to the example of working with unstructured data in a pdf to extract structured data.