Class ConfigFastHessian

java.lang.Object
boofcv.abst.feature.detect.interest.ConfigFastHessian
All Implemented Interfaces:
Configuration, Serializable

public class ConfigFastHessian
extends Object
implements Configuration
Configuration for FastHessianFeatureDetector plus feature extractor.
See Also:
Serialized Form
  • Field Details

    • extract

      public ConfigExtract extract
      Configuration for non-maximum thresholding
    • maxFeaturesPerScale

      public int maxFeaturesPerScale
      The maximum number of features it can detect in a single scale. Useful if you want to prevent high frequency features from dominating. If ≤ 0 then it will have no limit.
    • maxFeaturesAll

      public int maxFeaturesAll
      Maximum number of features it will return in total. If ≤ 0 then it will have no limit.
    • selector

      public ConfigSelectLimit selector
      Approach used to select features when more than the maximum have been detected
    • initialSampleStep

      public int initialSampleStep
      How often pixels are sampled in the first octave. Typically 1 or 2.
    • initialSize

      public int initialSize
      Typically 9.
    • numberScalesPerOctave

      public int numberScalesPerOctave
      Typically 4.
    • numberOfOctaves

      public int numberOfOctaves
      Typically 4.
    • scaleStepSize

      public int scaleStepSize
      Increment between kernel sizes as it goes up in scale. In some data sets, increasing this value beyound the default value results in an improvement in stability. Default 6
  • Constructor Details

    • ConfigFastHessian

      public ConfigFastHessian​(float detectThreshold, int extractRadius, int maxFeaturesPerScale, int initialSampleStep, int initialSize, int numberScalesPerOctave, int numberOfOctaves)
    • ConfigFastHessian

      public ConfigFastHessian()
  • Method Details