Skip to main content

Table 1 Description of phenotype texture measurements

From: Hierarchical classification strategy for Phenotype extraction from epidermal growth factor receptor endocytosis screening

Feature name Expression Description
std \(f_{1}=\sqrt {\sum \limits _{i}(i-mean)^{2}H\,(i)}\) The standard deviation of
   intensity from all the pixels
   in a region.
Smoothness \(f_{2}=1-\frac {1}{(1+{f_{1}^{2}})}\) The relative smoothness of the
   intensity in a region. It is 0 for a
   region of constant intensity and
   1 for a region with large excursion
   in the values of its intensity levels.
Skewness \(f_{3}=\sum \limits _{i}(i-mean)^{3}H\,(i)\) The order moment about the
   mean. The departure from
   symmetry about the mean
   intensity. It is 0 for symmetric
   histograms, positive for
   histograms skewed to the right
   and negative for histograms
   skewed to the left.
Uniformity \(f_{4}=\sum \limits _{i}H^{2}\,(i)\) The sum of squared elements in
   Histogram. It reaches maximum
   when all intensity levels are equal
   and decreases from there.
Entropy \(f_{5}=-\sum \limits _{i}H(i)\log _{2}{H\,(i)}\) The statistical measure of
   randomness.
  1. i represents the intensity value
  2. H(i) is the histogram of intensity
  3. mean symbolizes the average intensity