1
$\begingroup$

Let's say there is a dataset consisting of photos of cars from various brands, and we're trying to train a ConvNet to identify the brand from a photo, just like these ones:

enter image description here

One approach I was thinking of is a two-stages approach:

  1. isolate, for each brand, some distinctive tell-tale landmarks (e.g. the double radiator grill for BMWs or the headlights for Porsches)
  2. pre-train the network with these photos only at first
  3. once the network gets good accuracy for the above, then train it on photos showing entire cars

The set of first-stage-training photos would look like the following:

enter image description here

In both stages, the output variables would be exactly the same (the car brands). The difference is only that we train with one kind of photos at first, and then use a different kind of photos.

Remember that the ultimate goal is that the network is able to tell a car brand from photos showing the entire car. We don't care about its ability to identify a particular headlight from another, it's just an intermediate step along the process.

Questions:

  • is this approach common for image classification?
  • is there any literature on the topic?
  • is it any useful or is it better to let the network learn by itself?
  • should we simply mix some feature-centered images together with entire car images in one single training stage instead of separating two stages, with similar results?
$\endgroup$

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.