Abstract
The histogram of an image is the accurate graphical representation of thenumerical grayscale distribution and it is also an estimate of the probabilitydistribution of image pixels. Therefore, histogram has been widely adopted tocalculate the clustering means and partitioning thresholds for imagesegmentation. There have been many classical histogram-based image segmentationmethods proposed and played important roles in both academics and industry. Inthis tutorial, the histories and recent advances of the histogram-based imagesegmentation techniques are first reviewed and then they are divided into fourcategories: (1) the means-based method, (2) the Gaussian-mixture-model-basedmethod, (3) the entropy-based method and (4) the feature-points-based method.The purpose of this tutorial is threefold: 1) to teach the principles of theclassical histogram-based image segmentation methods to the interested readers;2) to evaluate the advantages and disadvantages of these classicalhistogram-based image segmentation methods objectively; 3) to compare theperformances of these classical histogram-based image segmentation methods withstate-of-the-art deep learning based methods objectively.