GLSZM_global_features - cerr/CERR GitHub Wiki

The feature definitions are available in IBSI section 3.8

List of available features

1. Small Area Emphasis (SAE) emphasizes small zones. A higher value indicates fine texture.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\frac{\textbf{P}(i,j)}{j^2}}}{N_z}$

2. Large Area Emphasis (LAE) emphasizes large zones. A higher value indicates coarse texture.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\textbf{P}(i,j)j^2}}{N_z}$

3. Gray Level Non-Uniformity (GLN) measures the distribution of zone counts over the grey values. The feature value is low when zone counts are equally distributed along grey levels.

$\frac{\sum^{N_g}_{i=1}\left(\sum^{N_s}_{j=1}{\textbf{P}(i,j)}\right)^2}{N_z}$

4. Gray Level Non-Uniformity Normalized is a normalised version of the grey level non-uniformity feature.

$\frac{\sum^{N_g}_{i=1}\left(\sum^{N_s}_{j=1}{\textbf{P}(i,j)}\right)^2}{N_z^2}$

5. Size Zone Non-Uniformity (SZN) measures the distribution of zone counts over the different zone sizes. Zone size non-uniformity is low when zone counts are equally distributed along zone sizes.

$\frac{\sum^{N_s}_{j=1}\left(\sum^{N_g}_{i=1}{\textbf{P}(i,j)}\right)^2}{N_z}$

6. Size Zone Non-Uniformity Normalized normalised version of zone size non-uniformity.

$\frac{\sum^{N_s}_{j=1}\left(\sum^{N_g}_{i=1}{\textbf{P}(i,j)}\right)^2}{N_z^2}$

7. Zone Percentage (ZP) measures the fraction of the number of realized zones and the maximum number of potential zones. Highly uniform ROIs produce a low zone percentage.

$\frac{N_z}{N_p}$

8. Low Gray Level Zone Emphasis (LGLZE) is a grey level analogue to small zone emphasis. Instead of small zone sizes, low grey levels are emphasized.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\frac{\textbf{P}(i,j)}{i^2}}}{N_z}$

9. High Gray Level Zone Emphasis (HGLZE) is a grey level analogue to large zone emphasis. The feature emphasises high grey levels.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\textbf{P}(i,j)i^2}}{N_z}$

10. Small Area Low Gray Level Emphasis (SALGLE) emphasizes zone counts within the upper left quadrant of the GLSZM, where small zone sizes and low grey levels are located.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\frac{\textbf{P}(i,j)}{i^2j^2}}}{N_z}$

11. Small Area High Gray Level Emphasis (SAHGLE) emphasizes zone counts in the lower left quadrant of the GLSZM, where small zone sizes and high grey levels are located.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\frac{\textbf{P}(i,j)i^2}{j^2}}}{N_z}$

12. Large Area Low Gray Level Emphasis (LALGLE) emphasizes zone counts in the upper right quadrant of the GLSZM, where large zone sizes and low grey levels are located.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\frac{\textbf{P}(i,j)j^2}{i^2}}}{N_z}$

13. Large Area High Gray Level Emphasis (LAHGLE) emphasizes zone counts in the lower right quadrant of the GLSZM, where large zone sizes and high grey levels are located.

$\frac{\sum^{N_g}_{i=1}\sum^{N_s}_{j=1}{\textbf{P}(i,j)i^2j^2}}{N_z}$

14. Grey Level Variance measures the variance in zone counts for the grey levels.

$\displaystyle\sum^{N_g}_{i=1}\displaystyle\sum^{N_s}_{j=1}{p(i,j)(i - \mu)^2}$

15. Size Zone Variance measures the variance in zone counts for the different zone sizes.

$\displaystyle\sum^{N_g}_{i=1}\displaystyle\sum^{N_s}_{j=1}{p(i,j)(j - \mu)^2}$

where
is the total number of gray levels in the image.
is the total number of zone sizes in the image.
is the total number of voxels in the image.
is the number of zones in the image.
$\textbf{P}(i,j)$ is the size zone matrix.
$p(i,j) = \frac{\textbf{P}(i,j)}{N_z}$ is the normalized size zone matrix.
$\mu = \displaystyle\sum^{N_g}_{i=1}\displaystyle\sum^{N_s}_{j=1}{p(i,j)i}$

Calculation details

Let quantizedM be the discretized scan matrix. The number of voxels in the region of interest is calculated as

numVoxels = sum(~isnan(quantizedM(:)));

Now, let's build the size zone matrix. The size zone matrix can be built using 2-d or 3-d zones. The szmType argument to the calcSZM function controls this.

szmType = 1; % 1: 3d, 2: 2d

The size zone matrix is computed as follows:

szmM = calcSZM(quantizedM, numGrLevels, szmType);

szmM is of size `(numGrLevels x numZoneSizes)

Generate flags for individual features

szmFeatC = {'sae',    'lae',    'gln',    'glv',    'szv',    'glnNorm',    'szn',    'sznNorm',...
    'zp',    'lglze',    'hglze',    'salgle',    'sahgle',    'lalgle',    'larhgle'};
szmFlagS = cell2struct(num2cell(ones(size(szmFeatC ))),szmFeatC,2);

Finally, the scalar features from the SZM matrix are obtained as follows:

featureS = szmToScalarFeatures(szmM,numVoxels,szmFlagS);

GLSZM_global_features - cerr/CERR GitHub Wiki

List of available features

Calculation details

⚠️ **GitHub.com Fallback** ⚠️

⚠️ GitHub.com Fallback ⚠️