The original plan for this project was to classify a receipt as either a receipt or not a receipt. An update to this is to classify the sections of a receipt e.g. the logo, items, footer etc. This should overall make this more interesting.
Currently I have a datase of images.
------ Update
I have started to create a neural network without the use of TF or other ML packages. This project is not currently being worked on but be revisited in the future when I have more knowledge.