Module: Auto Thresholding / Auto Threshold High ()

Description:

Auto Threshold High is one of the configurations of the module Auto Thresholding. For other configurations, see the port Type.

This module computes an automatic threshold on a grayscale image i.e., separate the image in 2 classes of pixels. Four methods of classification are available: Entropy, Factorization, Moments and IsoData. The computed threshold is displayed in the Tables Panel.

Entropy
The entropy principle defines 2 classes in the image histogram by minimizing the total classes' entropy, for more theory the reader can refers to references [1] and [2]. Considering the first-order probability histogram of an image and assuming that all symbols in the flowing equation are statistically independent, its entropy (in the Shannon sense) is defined as:



Where is the number of grayscales, the probability of occurrence of level and is the log in base 2.

Let us denote the value of the threshold and the search interval. We can define two partial entropies:





Where defines the probability of occurrence of level in the range and defines the probability of occurrence of level in the range. We search the threshold value which minimizes the sum :



Figure 1: Example of thresholding using the entropy method

Factorization
The factorization method is based on the Otsu criterion (see [3] for details), i.e., minimizing the within-class variance:



Where and are respectively the probabilities occurrence and , the variances of classes and .

A faster and equivalent approach is to maximize the between-class variance:



The within-class variance calculation is based on the second-order statistics (variances) while the between-class variance calculation is based on the first order statistics (means). It is, therefore, simplest and faster to use this last optimization criterion. We then search the value which maximizes the between-class variance such as:



Figure 2: Example of thresholding using the factorization method

Moments
The moment Auto Threshold High uses the moment-preserving bi-level thresholding described by W.H.Tsai in [4]. Moments of an image can be computed from its histogram in the following way:



Where is the probability of occurrence of grayscale . For the following we note the original grayscale image and the threshold image. Image can be considered as a blurred version of an ideal bi-level image which consists of pixels with only two gray values: and . The moment-preserving thresholding principle is to select a threshold value such that if all below- threshold gray values of the original image are replaced by and all above threshold gray values replaced by , then the first three moments of the original image are preserved in the resulting bi-level image. Image so obtained may be regarded as an ideal unblurred version of . Let and denote the fractions of the below-threshold pixels and the above-threshold pixels in , respectively, then the first three moments of are:



And preserving the first three moments in , means the equalities:



To find the desired threshold value , we can first solve the four equations system to obtain and , and then choose as the -tile of the histogram of . Note that and will also be obtained simultaneously as part of the solutions of system.

Figure 3: Example of thresholding using the moment-preserving method

IsoData
The IsoData Auto Threshold implements an iterative global thresholding algorithm which is based on the gray value histogram of the data.

References
[1] T.Pun, Entropic thresholding: A new approach, comput. Graphics Image Process. 16, 1981, 210-239
[2] J. N. Kapur, P. K. Sahoo, and A. K. C. Wong, "A New Method for Gray- Level Picture Thresholding Using the Entropy of the Histogram"; Computer Vision, Graphics and Image Processing 29, pp. 273-285, Mar. 1985
[3] Otsu, N. 1979. A thresholding selection method from grayscale histogram. IEEE Transactions on Systems, Man, and Cybernetics9(1): 62-66
[4] Tsai, W. H. 1985. Moment-preserving thresholding: A New Approach. Computer Vision, Graphics, and Image Processing 29: 377-393

See also: Adaptive Thresholding.

Connections:

Input Image [required]
The image to be thresholded. Supported types include: Grayscale images except float (Uniform Scalar Field).

Ports:

Type

This port allows selecting the configuration of this module. The available configurations are:

Interpretation

This port specifies whether the input will be interpreted as a 3D volume or a stack of 2D images for processing.

Mode

This port allows defining the following modes for auto thresholding computation:

Note: this port is visible if auto segment 3 phases is selected.

Input Range

This port defines a range where the threshold level will be defined by the chosen criterion before being used on the whole data range following the chosen configuration.

For example, it is useful if your data contains a lot of noise on one value: if you exclude this value from the range the value will be ignored from the threshold level definition and can improve your result.

Note: this port is unused if the port Mode is set to min-max.

Criterion

This option refers to the measure of dispersion used in the algorithm. The variance yields better results in most cases.
Entropy: Entropy of the intensity distribution.
Factorisation: Variance of the intensity distribution.
Moments: Moments of the intensity distribution.
IsoData: Iterative global thresholding algorithm which is based on the gray value histogram of the data.