How is image detection working without sliding windows?

In the 2nd week, we learn about sliding windows and other methodologies. But how does the assignment in the first week work? Does it use sliding windows method too? How is the bounding boxes determined in that method? It seemed fast to me as well.

Hey @Kirang96, Tim here. To answer your question, yes it does. The bounding boxes are determined via the window size and the feature map is an effect of the kernel convolution.

1 Like