equ_histo_image_rect🔗
Short description🔗
equ_histo_image_rect — Histogram linearization within a rectangluar mask.
Signature🔗
equ_histo_image_rect( image Image, out image ImageEquHisto, string Mode, extent.x MaskWidth, extent.y MaskHeight, real MaxContrast )
Description🔗
The operator equ_histo_image_rect enhances the contrast.
Similar to equ_histo_image, it applies a transformation to linearize
the cumulative histogram. However, instead of using the histogram of the
whole image, only the local neighborhood of each pixel is considered to
compute the transformation. The size of this rectangular neighborhood region
can be set by MaskWidth and MaskHeight.
The local contrast maximization improves the visibility of low-contrast
structures, but also amplifies noise. The parameter MaxContrast can
be used to limit the maximum contrast in a local neighborhood, which
effectively reduces the amplification of noise. Therefore, the method is also
known as Contrast-Limited Adaptive Histogram Equalization (CLAHE).
The parameter Mode determines the processing mode.
In the 'accurate' mode, the transformation is computed for each
pixel as described above.
In the 'fast' mode, the transformation is computed only for a subset
of all pixels and interpolated between those points.
Due to the interpolation, the results of both modes can differ significantly.
Attention🔗
The operator equ_histo_image_rect primarily serves for
optical processing of images for a human viewer. For example, the
contrast spreading can lead to a detection of fictitious edges.
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)
Image to be enhanced.
ImageEquHisto (output_object) (multichannel-)image(-array) → object (byte)
Image with linearized gray values.
Mode (input_control) string → (string)
Processing mode.
Default: 'accurate'
Suggested values: 'accurate', 'fast'
MaskWidth (input_control) extent.x → (integer)
Width of the filter mask.
Default: 51
Suggested values: 31, 51, 101, 151
Value range: 1 ≤ MaskWidth
Minimum increment: 2
Restriction: MaskWidth <= width(Image)
MaskHeight (input_control) extent.y → (integer)
Height of the filter mask.
Default: 51
Suggested values: 31, 51, 101, 151
Value range: 1 ≤ MaskHeight
Minimum increment: 2
Restriction: MaskHeight <= height(Image)
MaxContrast (input_control) real → (real / integer)
Maximum contrast.
Default: 0.01
Suggested values: 0, 0.01, 0.02, 0.05, 0.1, 0.5, 1
Value range: 0 ≤ MaxContrast ≤ 1
Complexity🔗
In Mode='accurate', equ_histo_image_rect uses an
algorithm with constant complexity per pixel, i.e., the runtime only depends
on the size of the input image and not on the mask size.
In Mode='fast', the number of histograms to be computed
depends on the mask size. Hence, the runtime increases with smaller mask
sizes and decreases with larger mask sizes.
Combinations with other operators🔗
Combinations
Possible successors
Alternatives
equ_histo_image, scale_image, scale_image_max, illuminate
See also
References🔗
S. Pizer et al.: “Adaptive Histogram Equalization and Its Variations”; Computer Vision, Graphics, and Image Processing, vol. 39, no. 3, pp. 355-368, 1987, doi: 10.1016/S0734-189X(87)80186-X.
Philipp Härtinger, Carsten Steger: “Adaptive histogram equalization in constant time”; Journal of Real-Time Image Processing, vol. 21, article 93, 2024, doi: 10.1007/s11554-024-01465-1.
Module🔗
Foundation