You are viewing the course site for a past offering of this course. The current offering may be found here.
Lecture 22: Image Processing (23)
glinia

Is there a reason Canny edge detection is commonly used as opposed to this scheme? Both seem equally effective. (As a side note, the edge-detected picture looks like an artist drew it!)

tyleryath

Do convolutional neural nets attempt to do things like edge detection by using convolutional layers? Is that similar to this form of edge detection or are they completely unrelated?

anup-h

@tylerath the Sobel edge detector in this picture uses convolution with a filter as a part of the algorithm, so it is very well suited to be implemented using a convolutional neural net

anup-h

@glinia The Canny edge detector is a more in-depth, more complicated algorithm than simply convolving two filters; it has steps to perform edge thinning, so that unlike the example on this slide, only one edge is preserved for sharp gradients in the original picture, making it far better suited for object recognition and other tasks. https://docs.opencv.org/master/da/d22/tutorial_py_canny.html

You must be enrolled in the course to comment