auto_threshold — Segment an image using thresholds determined from its histogram.
auto_threshold segments a single-channel image using multiple thresholding. First, the absolute histogram of the gray values is determined. Then, relevant minima are extracted from the histogram, which are used successively as parameters for a thresholding operation. The thresholds used for byte images are 0, 255, and all minima extracted from the histogram (after the histogram has been smoothed with a Gaussian filter with standard deviation Sigma). For each gray value interval one region is generated. Thus, the number of regions is the number of minima + 1. For uint2 images, the above procedure is used analogously. However, here the highest threshold is 65535. Furthermore, for uint2 images the value of Sigma (virtually) refers to a histogram with 256 values, although internally histograms with a higher resolution are used. This is done to facilitate switching between image types without having to change the parameter Sigma. For float images the thresholds are the minimum and maximum gray value in the image and all minima extracted from the histogram. Here, the scaling of the parameter Sigma refers to the original gray values of the image. The larger the value of Sigma is chosen, the fewer regions will be extracted. This operator is useful if the regions to be extracted exhibit similar gray values (homogeneous regions).
Regions with gray values within the automatically determined intervals.
Sigma for the Gaussian smoothing of the histogram.
Default value: 2.0
Suggested values: 0.0, 0.5, 1.0, 2.0, 3.0, 4.0, 5.0
Typical range of values: 0.0 ≤ Sigma ≤ 100.0 (lin)
Minimum increment: 0.01
Recommended increment: 0.3
Restriction: Sigma >= 0.0
read_image (Image, 'fabrik') median_image (Image, Median, 'circle', 3, 'mirrored') auto_threshold (Median, Seg, 2.0) connection (Seg, Connected)
anisotropic_diffusion, median_image, illuminate
connection, select_shape, select_gray
gray_histo, gray_histo_abs, histo_to_thresh, smooth_funct_1d_gauss, threshold