Scribd
Upload
31 views5 pages

IJ-801-An Image Filter For Counting Pixel Neighbors

Uploaded by

Rifky pramayandi mahesa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views5 pages

IJ-801-An Image Filter For Counting Pixel Neighbors

Uploaded by

Rifky pramayandi mahesa
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 5

An Image Filter for Counting Pixel Neighbors

Release 0.01

Robert Tamburo1

May 24, 2011


1 [email protected]

Abstract

This papers describes an image filter that counts the number of neighbors a pixel has storing that value at
the pixel’s image index. Functionality is provided to 1) adjust the neighborhood size, 2) count only those
pixels within a specific value range, and 3) only record a neighbor count for specific pixels of interest.
This paper is accompanied with source code for the filter and test, test images and parameters, and
expected output images.

Latest version available at the Insight Journal [ http://hdl.handle.net/10380/3267]


Distributed under Creative Commons Attribution License

Contents
1 Implementation of Algorithm 1

2 Example Usage 2

3 Software Used 2

1 Implementation of Algorithm

CountNeighborsImageFilter counts the number of neighbors a pixel has and stores the count at the pixel’s
image index. The entire is image is iterated over and at each pixel a neighborhood, centered at the current
pixel, is constructed. Pixels within the neighborhood are counted if the user-defined criteria is satisfied.
The count does not include the pixel of interest. The size of the neighborhood can be set with the function
SetRadius(), which has a default setting of 1 to contain all fully connected pixels. Several options for
count inclusion criteria are available:

• SetCountAtValue(): Counts all pixels equivalent to specified value.


2

• SetCountAboveValue(): Counts all pixels above a specified value.

• SetCountBelowValue(): Counts all pixels between a specified value.

• SetCountBetweenValue(): Counts all pixels between two specified values.

• SetCountNonZero(): Counts all non-zero pixels (default).

SetValueOfInterest() allows the user to only count neighbors for pixels of a specific value, which is
disabled by default

2 Example Usage

This filter requires setting an input image SetInput(), the radius of the neighborhood with SetRadius(),
the counting strategy, and the option to count neighbors for specific pixels. Example usage is shown below.

typedef itk::CountNeighborsImageFilter<ImageType, ImageType> FilterType;


FilterType::Pointer filter = FilterType::New();
filter->SetInput(inputImage);
ImageType::SizeType radius;
radius.Fill(1);
filter->SetRadius(radius); // set radius of neighborhood
filter->SetCountNonZero(1); // set counting strategy
filter->SetValueOfInterest(255); // set pixel of interest
filter->Update(); // execute filter

itkCountNeighborsImageFilterTest.cxx contains test examples with 4 different sets of parameters for


an input image (Fig. 1). The input image contains hand drawn patterns of value 127, 195, and 255.
The first set of parameters counts all non-zero pixels within a neighborhood of radius one and centered
around pixels with a value of 255 (Fig. 2).
The second set of parameters counts all pixels above 254 within a neighborhood of radius one and centered
around any pixel (Fig. 3).
The third set of parameters counts all pixels between 126 and 196 within a neighborhood of radius two and
centered around any pixel (Fig. 4).
The fourth set of parameters counts only pixels with a value of 255 within a neighborhood of radius one and
centered around pixels with a value of 127 (Fig. 5).

3 Software Used

This filter was developed on a Windows 7 64-bit computer. It has been successfully tested with ITK version
3.18.0, MinGW version 5.1.6, and CMake version 2.8.2 (Windows binary).

Latest version available at the Insight Journal [ http://hdl.handle.net/10380/3267]


Distributed under Creative Commons Attribution License
3

Figure 1: The input image input.png for this example.

Figure 2: The output image produced with the first set of parameters. Intensities scaled in a viewer for visualization
purposes

Latest version available at the Insight Journal [ http://hdl.handle.net/10380/3267]


Distributed under Creative Commons Attribution License
4

Figure 3: The output image produced with the second set of parameters. Intensities scaled in a viewer for visualization
purposes

Figure 4: The output image produced with the third set of parameters. Intensities scaled in a viewer for visualization
purposes

Latest version available at the Insight Journal [ http://hdl.handle.net/10380/3267]


Distributed under Creative Commons Attribution License
5

Figure 5: The output image produced with the fourth set of parameters. Intensities scaled in a viewer for visualization
purposes

Latest version available at the Insight Journal [ http://hdl.handle.net/10380/3267]


Distributed under Creative Commons Attribution License

You might also like