Suppose that we have a labeled training set of $n$ closely cropped images of cars $(x_1, y_1) , \dots, (x_n, y_n)$. We then train a CNN on this. Let's say we have $m$ test images. Then for each of the $m$ images, do we use the trained CNN on a cropped out portion of the box to detect whether there is a car or not? If the object is large, wouldn't having a large sliding window have better performance than a smaller sliding window?

  • $\begingroup$ Hi and welcome to this community! I think it would be useful if you can explain more in detail the relationship between the sliding window and the test dataset: why do you need a sliding window in the first place? Are you referring to a specific model that uses a sliding window? If yes, edit your question to add these details. What is the context of your question, model and problem? $\endgroup$ – nbro Sep 9 '19 at 22:00
  • $\begingroup$ @nbro: pyimagesearch.com/2015/03/23/… The idea is to use the "non-convolutional" version of YOLO. $\endgroup$ – ekjrnke Sep 9 '19 at 23:42

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.