What you need to complete the steps below to build the neural network (For recognition text from image)?

Keras Examples. In a lot of good examples on popular datasets. Explore, analyze, apply.

There's a wonderful book Tariq Rashid, "Make your own neural network". Tells step by step how to write that same network for recognition of images, slang is not for PhDs, but for ordinary mortals.
It is advisable to read the original, it will be much clearer even if not very strong in English.