Top  Previous  Next

Menu:    Processing > Segmentation…  

Use this command to perform segmentation of images. Segmentation is a process by which objects of interest in an image are separated from the rest of the image. The regions corresponding to the objects of interest are referred as the foreground, while the rest is referred as background.

In general, the segmentation produces the image of the multiphase type where each pixel value represents a separate class (phase) of objects. A particular case of the multiphase type is the binary image with only two phases present. Pixels whose values are equal to 1 correspond to the foreground objects, while pixels with zero values represent the background.

The most traditional method of segmentation is thesholding which is accomplished by defining a range of pixel values that correspond to the foreground or each foreground phase. Thresholds could be set interactively by a user or obtained automatically by analyzing the image histogram. A more advanced thresholding technique use the adaptive histogram analysis applying different thresholds to different areas of the image.

ImageWarp incorporates a number of interactive and automatic segmentation functions that can be applied to grayscale images as well as to color ones. Select one of the following functions:


Binary Threshold
perform manual binary thresholding of a grayscale image.
Multiphase Threshold
perform manual multiphase thresholding of a grayscale image.
RGB Threshold
perform manual RGB thresholding of a color image.
HLS Threshold
perform manual HLS thresholding of a color image.



Area percent
perform automatic thresholding based on the relative area of objects of interest.
Trim Histogram
perform automatic thresholding by trimming the histogram of an image.
Robust Unimodal
perform automatic thresholding based on the unimodal distribution.
Minimum Search
perform automatic thresholding based on the histogram minimum search.
Minimize Error
perform automatic thresholding by minimizing the distribution error.
Minimize Variance
perform automatic thresholding based on the minimum of variance.
Gaussian Fit
perform automatic threshodling by fitting the Gaussian distribution to histogram.
perform automatic thresholding by finding the valleys in the histogram.



Gaussian Fit
perform adaptive segmentation by fitting the Gaussain to the local histogram.
Minimize Variacne
perform adaptive segmentation based on the local minimum of variance.
Minimize Error
perform adaptive segmentation by minimizing the local distribution error.