eliminate_min_max🔗
Short description🔗
eliminate_min_max — Smooth an image in the spatial domain to suppress noise.
Signature🔗
eliminate_min_max( image Image, out image FilteredImage, extent.x MaskWidth, extent.y MaskHeight, number Gap, integer Mode )
Description🔗
eliminate_min_max smooths an image by replacing gray values with
neighboring mean values, or local minima/maxima. In order to prevent
edges and lines from being smoothed, only those gray values that represent
local minima or maxima are replaced (if there is a line or edge within an
image there will be at least one neighboring pixel with a comparable
gray value). Gap controls the strictness of replacement: Only
gray values that exceed all other values within their local neighborhood
more than Gap and all values that fall below their neighboring
more than Gap are replaced.
Thus, with \(g(x,y)\) being the gray value at position \((x,y)\) and \(E(x,y)\) representing the gray values of a \(\textrm{MaskWidth} \times \textrm{MaskHeight}\) sized rectangular neighborhood of a pixel at position \((x,y)\), containing all pixels within the neighborhood except the pixel itself, a pixel is replaced
-
if \(g(x,y) \ge \textrm{Gap} + max(E(x,y))\), or
-
else if \(g(x,y) + \textrm{Gap} <= min(E(x,y))\).
-
Elsewise \(g(x,y)\) is adopted without change.
Mode specifies how to perform the new value in case of a
replacement.
-
Mode= 1: replace a local maximum with next minor local maximum and replace a local minimum with next bigger local minimum. -
Mode= 2: replace with mean value of all pixels within the local neighborhood (including the replaced pixel). -
Mode= 3: replace with median value of all pixels within the local neighborhood (including the replaced pixel (also used ifModehas got any other value than 1 or 2).
MaskWidth and MaskHeight specify the width and height of
the rectangular neighborhood.
Border treatment: Pixels outside the image border are not considered (e.g.,
with a local \(3 \times 3\)-mask the neighborhood of a pixel at
\((0,0)\) reduces to the pixels at \((1,0)\),
\((0,1)\), and \((1,1)\)).
For an explanation of the concept of smoothing filters see the introduction of chapter Filters / Smoothing.
Attention🔗
If MaskWidth or MaskHeight is an even
number, it is replaced by the next higher odd number (this allows the
unique extraction of the center of the filter mask).
Width/height of the mask may not exceed the image width/height.
Note that filter operators may return unexpected results if an image with a reduced domain is used as input. Please refer to the chapter Filters.
Execution information🔗
Execution information
-
Multithreading type: reentrant (runs in parallel with non-exclusive operators).
-
Multithreading scope: global (may be called from any thread).
-
Automatically parallelized on tuple level.
-
Automatically parallelized on channel level.
-
Automatically parallelized on domain level.
Parameters🔗
Image (input_object) (multichannel-)image(-array) → object (byte / uint2)
Image to smooth.
FilteredImage (output_object) (multichannel-)image(-array) → object (byte / uint2)
Smoothed image.
MaskWidth (input_control) extent.x → (integer)
Width of filter mask.
Default: 3
Suggested values: 3, 5, 7, 9
Value range: 3 ≤ MaskWidth ≤ width(Image)
Minimum increment: 2
Recommended increment: 2
Restriction: odd(MaskWidth)
MaskHeight (input_control) extent.y → (integer)
Height of filter mask.
Default: 3
Suggested values: 3, 5, 7, 9
Value range: 3 ≤ MaskHeight ≤ width(Image)
Minimum increment: 2
Recommended increment: 2
Restriction: odd(MaskWidth)
Gap (input_control) number → (real)
Gap between local maximum/minimum and all other gray values of the neighborhood.
Default: 1.0
Suggested values: 1.0, 2.0, 5.0, 10.0
Mode (input_control) integer → (integer)
Replacement rule.
Default: 3
List of values: 1, 2, 3
Result🔗
eliminate_min_max returns 2 (H_MSG_TRUE) if all parameters are
correct. If the input is empty eliminate_min_max returns with
an error message.
Combinations with other operators🔗
Combinations
Possible successors
wiener_filter, wiener_filter_ni
See also
mean_sp, mean_image, median_image, median_weighted, binomial_filter, gauss_filter, smooth_image
References🔗
M. Imme:“A Noise Peak Elimination Filter”; S. 204-211 in CVGIP Graphical Models and Image Processing, Vol. 53, No. 2, March 1991
M. Lückenhaus:``Grundlagen des Wiener-Filters und seine Anwendung in der Bildanalyse’‘; Diplomarbeit; Technische Universität München, Institut für Informatik; Lehrstuhl Prof. Radig; 1995.
Module🔗
Foundation