While traditional smoothing methods use the same window shape and size to smooth an image independent of its
content, in adaptive smoothing, the window shape and size are varied across the image domain depending on local image
content. In adaptive smoothing a window is sized according to the local gradient magnitude and shaped in such a way that it
has a shorter side across an edge compared to along the edge. This mechanism maintains edge details while smoothing
random noise.
Adaptive smoothing reduces noise similar to traditional smoothing but unlike traditional smoothing it does not blur the
edges (as much). Examples of adaptive smoothing are given below.
content, in adaptive smoothing, the window shape and size are varied across the image domain depending on local image
content. In adaptive smoothing a window is sized according to the local gradient magnitude and shaped in such a way that it
has a shorter side across an edge compared to along the edge. This mechanism maintains edge details while smoothing
random noise.
Adaptive smoothing reduces noise similar to traditional smoothing but unlike traditional smoothing it does not blur the
edges (as much). Examples of adaptive smoothing are given below.
Fig. 1. (Left) An image corrupted by impulse noise. (Center) Traditional median filtering. (Right) Adaptive median filtering. The two
results look alike at first glance, but when studied more closely, small detailed appear to be preserved by the adaptive method but
removed by the traditional method.
results look alike at first glance, but when studied more closely, small detailed appear to be preserved by the adaptive method but
removed by the traditional method.
Fig. 2. (Left) An image corrupted by impulse noise. (Middle) Traditional median filtering. (Right) Adaptive median filtering. small
details are maintained by the adaptive method but smoothed by the traditional method.
details are maintained by the adaptive method but smoothed by the traditional method.
Fig. 3. (Top left) Image corrupted by white noise. (Top middle) Image smoothed by traditional mean filtering. (Top right) Image
smoothed by adaptive mean filtering. (Bottom left) Image smoothed by traditional Gaussian filtering. (Bottom right) Image
smoothed by adaptive Gaussian filtering. The difference between new and traditional methods should be obvious.
smoothed by adaptive mean filtering. (Bottom left) Image smoothed by traditional Gaussian filtering. (Bottom right) Image
smoothed by adaptive Gaussian filtering. The difference between new and traditional methods should be obvious.
Fig. 4. (Top left) Image corrupted by white noise. (Top middle) Image smoothed by traditional mean filtering. (Top right) Image
smoothed by adaptive mean filtering. (Bottom left) Image smoothed by traditional Gaussian filtering. (Bottom right) Image smoothed
by adaptive Gaussian filtering.
smoothed by adaptive mean filtering. (Bottom left) Image smoothed by traditional Gaussian filtering. (Bottom right) Image smoothed
by adaptive Gaussian filtering.
To obtain a software license for this adaptive smoothing, please follow the link =>
 |  |
| Adaptive smoothing |
| Image Registration and Fusion Systems |
