IM_THRESHOLD

Fixed mapping thresholding images (DIP_Image)

   B = IM_THRESHOLD(A,TYPE,PAR,INV)
   B = A*IM_THRESHOLD([],TYPE,PAR,INV)
   B = A*IM_THRESHOLD(TYPE,PAR,INV)

Input
A	Dataset with object images (possibly multi-band)
TYPE	Type of procedure, see below
PAR	Related parameter
INV	If INV = 1, result inverted, default INV = 0.

Output
B	Dataset with thresholded images

Description

The following procedures are supported (TYPE)

'isodata':	Thresholding using the Isodata algorithm
'otsu':	See IM2BW
'triangle':	*Thresholding using chord method  (a.k.a. skewed bi-modality, maximum distance to triangle)  by Zack, Rogers and Latt (1973).
'background':	*Thresholding using unimodal background-symmetry method.
'fixed':	Thresholding at a fixed value.
'double':	*Thresholding between two fixed values.
'volume':	*Thresholding to obtain a given volume fraction.
'hysteresis':	*From the binary image (in>low) only those regions are  selected for which at least one pixel is (in>high)

The following parameters are related to these procedures (PAR)
('background'): Distance to the peak where we cut-off, in  terms of the half-width at half the maximum.  Inf selects the default value, which is 2.

('fixed'):	Threshold value. Inf means halfway between  minimum and maximum value.
('double'):	Two threshold values. Inf means min+[1/3,2/3]*(max-min).
('volume'):	Parameter = the volume fraction (Inf means 0.5)
('hysteresis'):	Two values: low, high (see above)

* routine needs DIP_IMAGE

This routine is for smart thresholding. Simple thresholding can also be  performed by B = A > PAR, which is the same as using the type 'fixed'.

See also

datasets, datafiles, dip_image, threshold,