3

In image processing, how do region growing and clustering differ from each other? There are many papers like clustering based on region growing. I couldn't able to draw a fine line between these two concepts.

Phonon
  • 5,216
  • 5
  • 37
  • 62
Premnath D
  • 1,011
  • 2
  • 14
  • 25
  • Is this a homework question? – Phonon Feb 26 '14 at 00:21
  • no it is not.. I was reading region growing 2 days back, now while reading clustering techniques (k-means), it also looks similar to me. In region growing, we have seed points. In k-means clustering, we have means values to start with. So I would like to know the difference – Premnath D Feb 26 '14 at 00:25
  • Have you Googled it? – Phonon Feb 26 '14 at 00:34
  • yes, I did. It tells that region growing comes under clustering. There are many papers like clustering based on region growing. I couldn't able to draw a fine line between these two concepts. That is why asked more information on how they differ from each other – Premnath D Feb 26 '14 at 00:37
  • K-means does not take into account the spatial distribution of, for example, pixels in an image. Region growing does since it looks for nearby similar pixels. – KAE May 23 '18 at 13:27

1 Answers1

0

Clustering is a more general operation not only related to image pixels.
While region growing is operation of expanding area of image.

Both methods can be used for Image Segmentation.
The main difference being that region growing is local first operation.
Namely it labels together only pixels which are local.

Clustering on image pixels requires us to define the metric and features.
If we use as a feature the coordinates of the pixels and the values we can get something similar to region growing called Super Pixels (See Rectangle Segments of Image (Rectangle Super Pixels) per Pixel and Locate Non Homogeneous Areas in an Image).

Royi
  • 19,608
  • 4
  • 197
  • 238