Extract certain information from images based on predefined conditions

I am new in this area and I have a task to extract not all but certain information from tags like model, serial number. On tags there can be a lot of other information. Therefore I see this process like this a user scans an image, the AI/ML finds a similar image in the library(classification), recognizes all text elements and provides only a subset of fields linked to this type of image (model, serial number, producer etc). Which tools do I need to use to create a new model, train? How can a user create a new model and load picture and define a list of fields to be extracted?