Software can support adding extra metadata from embedded images, including photographs, graphs, maps and so on. An example approach is included here:
User decides to add metadata to a specific image, such as this one:
To do so the user chooses a ‘Add Metadata’ tool and draws a rectangle (potentially any shape) around the area in question:
The software then presents a dialog to the user (as made available by the software and as chosen by the user if there is more than one option). In this example the extraction of Metadata is done by third party AI software so the AI Software dialog, if any, would be displayed. To make this work feasible the AI may need a prompt to better inform what the image is and what should be extracted. Here is one such choice, in a very simplistic way:
Once verified by the user, the user can choose to click outside to dismiss or ‘Save’ to add a page of Visual-Meta with this extracted data.
As you can see below, when attempting this manually with ChatGPT at the end of 2024 it would need more focused prompts and possibly code:
How the Metadata should be added is available for experimentation.