Object recognition tools & commands?

Hello, everyone! I've got a task to implement an object recognition system. I have a training set of images and corresponding labels for them, and then the test set - to check the recognition rate. As far as I understand it should go as follows: extract features from training images, assign those features and labels to each class and then with some classification tool, sort the test set of images among those classes. The question is - what kind of tools would be more appropriate for my problem? I would very much appreciate any guidelines as to what approaches (or even specific commands) are better to choose, considering feature extraction and classification.

Here are some example images:

The training set and testing set look almost the same.

P.S. Sorry if the question is not very appropriate here, but I honestly couldn't find any clear advices on this matter on the internet.

Thanks in advance!