rank_image🔗
Short description🔗
rank_image — Compute a rank filter with arbitrary masks.
Signature🔗
rank_image( image Image, region Mask, out image ImageRank, integer Rank, string Margin )
Description🔗
rank_image performs a rank filter on the input image
Image with a mask that is specified by the region
Mask and returns the filtered image in ImageRank.
The shape of the mask can be chosen arbitrarily and can, for
example, be created with operators like gen_circle or
draw_region. The position of the mask does not influence
the result since the center of gravity of the mask region is used as
the reference point of the mask.
Conceptually, the rank filter sorts all gray values within the mask
in ascending order and then selects the gray value with rank
Rank. The rank 1 corresponds to the smallest gray value
and the rank \(A\) corresponds to the largest gray value within the
mask. Here, \(A\) denotes the area of Mask (see
area_center). For \(\textrm{Rank} = (A - 1) / 2 + 1\),
rank_image returns the median gray value (see
median_image). For \(\textrm{Rank} = 1\), rank_image
performs a gray value erosion (see gray_erosion_rect,
gray_erosion_shape, and gray_erosion), while for
\(\textrm{Rank} = A\) rank_image performs a gray value
dilation (see gray_dilation_rect,
gray_dilation_shape, and gray_dilation).
rank_image can be used, for example, to suppress noise or to
suppress unwanted objects that are smaller than the mask.
Furthermore, rank_image is less sensitive to noise than the
corresponding gray value morphology operators. Therefore, to obtain
a more robust version of the gray value morphology, instead of using
1 or \(A\), slightly larger or smaller values should be selected for
Rank.
Several border treatments can be chosen for filtering via the
parameter Margin:
-
gray value Pixels outside of the image border are assumed to be constant (with the indicated gray value).
-
'continued' Continuation of the gray values at the image border.
-
'cyclic' Cyclic continuation at the image borders.
-
'mirrored' Reflection of pixels at the image borders.
For an explanation of the concept of smoothing filters see the introduction of chapter Filters / Smoothing.
Attention🔗
rank_image uses an algorithm with a runtime per pixel that
depends on the number of runs in the mask Mask. Therefore,
rank_image is slower than rank_rect for rectangular
masks with a large mask height. The precise mask height for which
rank_rect will become faster than rank_image depends
on the computer architecture (processor type, availability of SIMD
instructions like SSE2 or MMX, cache size and throughput, memory
throughput). Typically, this is the case for mask heights \(>\) 15,
but can also be the case only for larger mask sizes, e.g., if SIMD
instructions are unavailable and memory throughput is low.
Furthermore, it should be noted that rank_rect uses a
recursive implementation, which internally computes the filter
response on the smallest enclosing rectangle of the domain of the
input image. Therefore, if the domain of the input image only
covers a small fraction of the smallest enclosing rectangle, it can
happen that rank_image is faster than rank_rect even
for larger mask heights.
rank_image should neither be used with \(\textrm{Rank} = 1\)
to perform a gray value erosion nor with \(\textrm{Rank} = A\) to
perform a gray value dilation. In these cases, the operators
gray_erosion_rect, gray_erosion_shape, or
gray_erosion and gray_dilation_rect,
gray_dilation_shape, or gray_dilation, respectively,
are typically faster than rank_image.
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 domain level.
Parameters🔗
Image (input_object) (multichannel-)image(-array) → object (byte / int2 / uint2 / int4 / real)
Image to be filtered.
Mask (input_object) region → object
Filter mask.
ImageRank (output_object) multichannel-image(-array) → object (byte / int2 / uint2 / int4 / real)
Filtered image.
Rank (input_control) integer → (integer)
Rank of the output gray value.
Default: 5
Suggested values: 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 31, 49, 51, 61, 71, 81, 91, 101
Value range: 1 ≤ Rank ≤ 4095
Minimum increment: 1
Recommended increment: 2
Margin (input_control) string → (string / integer / real)
Border treatment.
Default: 'mirrored'
Suggested values: 'mirrored', 'cyclic', 'continued', 0, 30, 60, 90, 120, 150, 180, 210, 240, 255
Example🔗
(HDevelop)
read_image (Image, 'fabrik')
draw_region (Region, WindowHandle)
rank_image (Image, Region, ImageRank, 5, 'mirrored')
dev_display(ImageRank)
Complexity🔗
For each pixel: \(O(N)\), where \(N\) is the number of runs of
Mask (see runlength_features).
Result🔗
If the parameter values are correct the operator rank_image
returns the value 2 (H_MSG_TRUE). The behavior in case of empty input (no
input images available) is set via the operator
set_system('no_object_result',<Result>). If necessary, an
exception is raised.
Combinations with other operators🔗
Combinations
Possible predecessors
read_image, draw_region, gen_circle, gen_rectangle1, gen_rectangle2
Possible successors
threshold, dyn_threshold, regiongrowing
Alternatives
rank_rect, median_image, median_rect
See also
gray_erosion_rect, gray_dilation_rect, gray_erosion_shape, gray_dilation_shape, gray_erosion, gray_dilation
References🔗
T.S. Huang, G.J. Yang, G.Y. Tang; ``A Fast Two-Dimensional Median Filtering Algorithm’‘; IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 27, no. 1, pp. 13-18, 1979.
R. Haralick, L. Shapiro; “Computer and Robot Vision”; Addison-Wesley, 1992, pp. 318-320.
Module🔗
Foundation