An example of what I'd like to do is identify the price of a product on a product page.

While I can train a CNN to identify prices, it's likely that it would recognise every instance of a price on a given page and end up picking up the prices of recommended products in addition to the price of the primary product.

Are there examples of visual neural architectures which take the context of the object into account? I'd imagine that would be something like a combination of an RNN and a CNN.

EDIT: Maybe a simpler approach would be to find a way to apply multiple labels to the same bounding box?

The reason I say this is that it if there are 10 prices on the page, it would be easier for the network to identify them first as prices, and secondarily apply a label for the primary price

  • $\begingroup$ I ain't sure about, but won't it work if we Split the network at the output end, as to identify the product as well as it's price... $\endgroup$
    – Aditya
    Jun 2 '18 at 1:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.