Histogram
← Thresholding | ● | Histogram Based Thresholding →
Thresholding is a sub case of image classification, which means that an image’s pixels needs to be classied as to which object they belong to. Thresholding is a classification into two classes: background and object.
Many classification algorithms (also called segmentation algorithms) depend on the so called histogram.
The histogram is a curve, which desribes the the frequency count of a particular scalar value.
For example a histogram of an image counts the occurances of the pixel values (gimp “Curves”):
Usually the brightness value of RGB pixels in the HSV color space is displayed as histogram curve, but any component of RGB, CMYK or HSV color spaces can serve as histogram basis.
As an example of a histogram of a scalar volume the histogram counts the occurances of the scalar values assigned to the voxels of the volume data set:
