First order (intensity) features - cerr/CERR GitHub Wiki

First order (intensity) features

These features are computed according to the definitions in IBSI section 4.3. The intensity-based statistical features describe how the intensity within the region of interest (ROI) is distributed. The features in this set do not require discretisation. Intensity-based statistical featuresare not meaningful if the intensity scale is arbitrary. Hence, normalization of intensities to a standard unit is required for modalities such as MR.

Following first order features are supported:

  • min:
  • max:
  • mean:
  • range:
  • std:
  • var:
  • median:
  • skewness:
  • kurtosis:
  • rms:
  • totalEnergy:
  • meanAbsDev:
  • medianAbsDev:
  • P10:
  • P90:
  • robustMeanAbsDev:
  • robustMedianAbsDev:
  • interQuartileRange:
  • coeffDispersion:
  • coeffVariation:

Example function call:

featureS.firstOrderS = radiomics_first_order_stats(scan3M(mask3M));