gray_insideGrayInsideGrayInsidegray_insidegray_inside🔗

Short description🔗

gray_insideGrayInsideGrayInsidegray_insidegray_inside — Calculate the lowest possible gray value on an arbitrary path to the image border for each point in the image.

Signature🔗

gray_inside( image Image, out image ImageDist )

HImage HImage::GrayInside( ) const

HImage HImage.GrayInside( )

Description🔗

gray_insideGrayInside determines the “cheapest” path to the image border for each point in the image, i.e., the path on which the lowest gray values have to be overcome. The resulting image contains the difference of the gray value of the particular point and the maximum gray value on the path. Bright areas in the result image therefore signify that these areas (which are typically dark in the original image) are surrounded by bright areas. Dark areas in the result image signify that there are only small gray value differences between them and the image border (which doesn’t mean that they are surrounded by dark areas; a small “gap” of dark values suffices). The value 0 (black) in the result image signifies that only darker or equally bright pixels exist on the path to the image border.

The operator is implemented by first segmenting into basins and watersheds the image using the watershedsWatersheds operator. If the image is regarded as a gray value mountain range, basins are the places where water accumulates and the mountain ridges are the watersheds. Then, the watersheds are distributed to adjacent basins, thus leaving only basins. The border of the domain (region) of the original image is now searched for the lowest gray value, and the region in which it resides is given its result values. If the lowest gray value resides on the image border, all result values can be calculated immediately using the gray value differences to the darkest point. If the smallest found gray value lies in the interior of a basin, the lowest possible gray value has to be determined from the already processed adjacent basins in order to compute the new values. An 8-neighborhood is used to determine adjacency. The found region is subtracted from the regions yet to process, and the whole process is repeated. Thus, the image is “stripped” form the outside.

Analogously to watershedsWatersheds, it is advisable to apply a smoothing operation before calling watershedsWatersheds, e.g., binomial_filterBinomialFilter or gauss_filterGaussFilter, in order to reduce the amount of regions that result from the watershed algorithm, and thus to speed up the processing time.

Attention🔗

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.

Parameters🔗

Imageimageimage (input_object) (multichannel-)image(-array) → object (byte)HObject (byte)HImage (byte)HObject (byte)Hobject (byte)

Image being processed.

ImageDistimageDistimage_dist (output_object) (multichannel-)image(-array) → object (int2)HObject (int2)HImage (int2)HObject (int2)Hobject * (int2)

Result image.

Example🔗

(HDevelop)

read_image(Image,'fabrik')
gauss_filter (Image,GaussImage,11)
gray_inside(GaussImage,ImageOut)
dev_display(ImageOut)

(C)

read_image(Image,"fabrik")\;
gauss_filter(Image,&GaussImage,11)\;
gray_inside(GaussImage,Result)\;

Result🔗

gray_insideGrayInside always returns 2 (H_MSG_TRUE).

Combinations with other operators🔗

Combinations

Possible predecessors

binomial_filterBinomialFilter, gauss_filterGaussFilter, smooth_imageSmoothImage, mean_imageMeanImage, median_imageMedianImage

Possible successors

select_shapeSelectShape, area_centerAreaCenter, count_objCountObj

See also

watershedsWatersheds

Module🔗

Foundation