4
$\begingroup$

What is difference between edge detection and contour detection?

Apparently both techniques are used to detect boundries of close shapes(wheter oval shape or r rectangular shape)

$\endgroup$
2
$\begingroup$

Contour is the edge closing an object. So you can think as higher level of edge detection.

So if an edge define an object it becomes a contour.

$\endgroup$
2
  • $\begingroup$ Consider a rectangle, it will have 4 edges(each edge across one side) and one contour made up of 4 edges? Am i right? $\endgroup$ – Man Apr 21 '20 at 11:45
  • 1
    $\begingroup$ Yep. This is the idea. $\endgroup$ – Royi Apr 21 '20 at 11:46
2
$\begingroup$

I would consider the notion of edge, with respect to that of contour, in the same relation as words to a sentence. An edge is a distinctive element of an image (a color change, a border cut, a shadow, an occlusion). Here, from top to bottom: surface normal, depeth, illumination, shadow.

Types of edges

But sometimes, this does not fully takes advantage of the whole concept of objects, which requires higher-order methods, semantics, viewpoints, etc. Look at the following 3D object, which is relatively homogeneous and isolated edges:

3D object, shadows and edges

While edges can be detected by relatively local tools (derivatives, Laplacien, thresholding, wavelets, Gabor functions etc.), contours often require more involved and global techniques: active contours, geodesic active gradients, etc.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.