Machine learning image analysis (category issue)

Hello!

I have questions about the machine learning.

What I'm trying to do is like recognizing images of dogs and cats (and other animals).

There are already a lot of tools for that, but I want to differentiate between the breeds below them.

For example, there are various breeds such as Golden Retriever, Chihuahua, Bulldog, etc.

So I want to make the sub-categories separate like dog - chihuahua and cat - persian.

There are my questions.

1. Is this possible? Could you tell me how to do it?

2. Can I do this with CNN (Convolutional Neural Network)?

Answer 1

[Image Analyst]

Yes, you need to do transfer learning. You need to follow the example for transfer learning and provide it with a labeled data set with hundreds of images of different breeds where you've correctly identified each breed. I'd probably say at least 20 of each breed (if the breeds are very distinctive/unique in their appearance) but more would of course be better.