Feature Extraction and use of background folder

[arghyadeep99]

With my vague understanding, what I comprehend is that we are just passing the images through a pre-trained network and then trying to perform image classification right? But what if we wanted to perform feature extraction to understand what exactly is the issue in each image and then say what kind of disease the plant is having?

Secondly, can you explain how exactly is the "background" folder is used from the train dataset folder used in your project? I understand why it's included, so that the model can comprehend real-world scenario or "background" for the leaf images right? But I am confused about how this folder is being utilized exactly in the project. Thank you.