I want to do the following project and I think the best way is using tensorflow/keras: I have photos as images from a list with prices of products and I need classify these text objects from the images such that i identify the product name and the price. So I think it is different from normal Image classification since my objects are text objects? First, I thought it is like object detection in images but I think i need more to take these texts to another neural network?
I think you are referring to a king of image auto captioning, and why not sentence alignement ! I suggest you to go over the paper Deep Visual-Semantic Alignments for Generating Image Descriptions which is one of the references today (by M. Karpathy). The first step is about recognizing objects in an image and align sentences to them (done with R-CNN and BRNN) and the second is about generating new captions (CNN + RNN)
Even though it may be quite difficult to deeply go through, there is some code available on GitHub and it has been remade by a a lot of people with various implementations. Here is the summary.
Good luck !