Barcodes classification project

Hi im trying to make a CNN project that detects Barcodes(QR/Datamatrix) in a pdf file or an image
I’m working the classification part now
I made a dataset that has QR code and Datamatrix code images and built a model but sometimes when i give the model a picture of QR/Datamatrix code it doesnt predict correctly what should i do to make the model better ?
ps: i used a data that have 720 images is it enough ?

Assuming that you’ve taken deep learning specialization, please clarify how this project is different from the ones covered in the courses.