Why we find Vertical and horizontal edges in Convolutional Neural Network?

Question

I'm working on a project of machine learning. One thing I'm not able to understand is that why we have to find vertical and horizontal edges in CNN?

Answer 1

We do so to separate the meaning full element from the whole data. For example in Image we will separate the part of image from the whole image to detect what it is this is done by detecting vertical and horizontal edges this will use to recognize the object within the image. If we have image of road and we need to find the specific car you would need to separate the cars from the whole image. So we need to detect vertical and horizontal edges and classify them into cars images to find out either they are cars or not. Hope this will clear you concept regarding the edges.