I was asked to verify the feasibility for solving a particular problem: recognizing for a fashion brand the model of its products.

I have little experience with image recognition in general, I always used Google Vision Api or some pre trained nets from Google/tensorflow like the VGG-16.

I am wondering: Do you think this level of granularity can be achieved?

I guess I should for example train the last layer of a pre-trained network, but I need many images to do that.


1 Answer 1



Do they have a unique feature on the item (e.g. a polo player, a red sole) the model can be trained on? I have seen people try identifying clothing model from images via Google's Vision API and it didn't work well. If you are in a fashion category where the business model is to copy designs like H&M or Zara do, you're likely going to have trouble.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.