Well it's your project so you need to decide which subject you need to understand first. Do you understand what "image classification" means in general, or do you have a specific idea about how you want to implement it? How do you think "pixel value ranges" will help you to classify images? And finally what do you know about the principles of artificial neural networks?
I hope that helps you to understand why such a question is impossible to answer in a forum such as this.
I posted an article in CodeProject and it got rejected because it has not been phrased in a way that allows it to be fully understood.
However, while I wrote the article, it was displayed correctly. It got screwed after publish. Please check the images below.
Before Publish: http://postimg.org/image/by2ens059/[^]
After Publish: http://postimg.org/image/n7ws1vzsl/[^]