``Relaxation_labeling''--Relaxation labeling using manually-set statistics of image regions

VFX name:

iRelaxLabeling

IMPROMPTU Equivalent:

Relaxation_labeling

purpose:

Uses relaxation labeling to segment an image into 2-4 distinct regions. The algorithms to compute the initial probability vectors, perform the updating, and compute a final segmentation can be found in [7] and [8]. Briefly, using the naming as is found in [8], the strategies used for this version of relaxation labeling are: initial probabilities--Basic, vector updating--fast general using support-reject compatibility coefficients. For final classification, each voxel gets the label corresponding to the largest vector component in that voxel's probability vector, provided it exceeds ``final_seg_thresh.''

Region parameters Mean $\mu$ and Standard Deviation $\sigma$ are manually set by user as opposed to Cue-based_Relaxation_Labeling where region parameters are computed from cues.

input:

An 8-bit gray scale volume.

output:

An 8-bit volume in segmentation volume format (refer to the Demask_Region function, p. ) containing the segmented regions.

parameters:

Default parameters menu is

number_regions = 2
Region 1 parameters:
first_mean: 0
first_std_dev: 0
Region 2 parameters:
second_mean: 0
second_std_dev: 0
Region 3 parameters:
third_mean: 0
third_std_dev: 0
Region 4 parameters:
fourth_mean: 0
fourth_std_dev: 0
num_iter = 10
final_seg_thresh = 0.000

If the final segmentation threshold is set to 0.0, then the maximum vector component for each voxel is assumed to represent that voxel's appropriate labeling. If a threshold is set, then the maximum vector component value for a voxel must be greater than the threshold value for that voxel to be labeled as part of a region.

comments:

See below.

1.

This ``turnkey'' version of the relaxation labeling function is described in detail in [8].

2.

Function works identically to Cue-based_Relaxation_Labeling, except that region parameters Mean $\mu$ and Standard Deviation $\sigma$ are set manually.