If I instruct my custom GPT to use Vision to read and analyse the images found in my uploaded PDF files, can GPT understand all the content, text along with the image illustrations?
Tried that with an PDF with embedded technical graphics, but it would not work and GPT told me it can not read the content of an image like a table or a graph.
Indeed, after asking GPT:
This task often involves specialized image recognition and OCR (Optical Character Recognition) technologies. It could be a developing area of AI that hasn’t been fully realized in a dedicated GPT yet
I wonder if it would be possible by using the Actions for calling some “image recognition” API…