Categories of Functions

Filter

Operations in the ``Filter'' category perform the basic image preprocessing functions of filtering and image enhancement. Some of the functions are linear, while many of the more effective functions are highly nonlinear. The list of functions here is:

Lowpass:

lowpass filter

Median:

median filter

Maximum_Homogeneity:

maximum-homogeneity filter

Symmetric_Nearest_Neighbor:

symmetric nearest-neighbor mean filter

Sigma:

sigma filter

Average_Maximum:

average/maximum filter

Average_Minimum:

average/minimum filter

Salt_and_Pepper_Removal:

remove spike noise

Anisotropic_Diffusion:

anisotropic diffusion

Morphology

Mathematical morphology is a popular framework for performing geometrically based operations on an image [19,25,3,21]. Since much of IMPROMPTU is geared toward extracting objects (or facilitating the eventual goal of extracting objects) and since geometrical properties are one of the richest class of attributes one can focus one when attempting to extract objects, mathematical morphology appears in many functions throughout IMPROMPTU. The ``Morphology'' category collects some of the more rudimentary (and more purely ``math-morphology-oriented'') functions.

The fundamental operations of erosion and dilation (and their gray-scale analogs of local minimum and local maximum) are, of course, included here. The complete list of functions here is:

Binary_Erosion:

binary erosion

Binary_Dilation:

binary dilation

Binary_Opening:

binary opening

Binary_Closing:

binary closing

Minimum:

local minimum (simple gray-scale erosion)

Maximum:

local maximum (simple gray-scale dilation)

Grayscale_Erosion:

gray-scale erosion; uses more complex structuring elements

Grayscale_Dilation:

gray-scale dilation; uses more complex structuring elements

Grayscale_Opening:

gray-scale opening; uses more complex structuring elements

Grayscale_Closing:

gray-scale closing; uses more complex structuring elements

Ultimate_Erosion:

ultimate erosion

Conditional_Dilation:

conditional dilation

Topology

A binary-valued volume (voxels take on the values 0 or 1) can be looked upon as a collection of connected sets (components), where a connected set consists of connected collections of 1's. Functions in the ``Topology'' category focus on the topological properties of such images. Examples of functions: finding (and labeling) connected components, deleting overly small components, deleting interior cavities, and thinning/thickening connected objects.

The list of functions here is:

3D_conn_comp:

three-dim. connected components analysis

2D_conn_comp:

two-dim. connected components analysis

Cavity_Deletion:

delete interior cavities

Homotopic_Thinning:

thin an input volume

Homotopic_Thickening:

conditional homotopic thickening

Segment_Generation:

compute more accurate central axes of thin structures

Segmentation

The category ``Segmentation'' contains functions that consist of integrated packages of simpler functions; a typical function in this category uses several simpler functions from the other categories. Most of these functions perform some sort of image segmentation operation.

The list of functions here is:

3D_Adaptive_thresholding:

automatically threshold a 3-D volume

hysteresis threshold:

-- thresholding using hysteresis; NOT IN IMPROMPTU, ROI ONLY!

Region_growing:

segment a volume using seeded region growing

Relaxation_Labeling:

image segmentation using cue-based relaxation labeling

Watershed_Relaxation:

image segmentation using cue-based, watershed-driven relaxation labeling

Watershed:

image segmentation using watershed analysis

Kirsch_edge_detection:

Kirsch edge operator

Difference_of_Gaussians:

Difference of Gaussians edge operator

Morphological_gradient:

morphological gradient

Binary_Border_detection:

find borders in a binary volume

Nonmaxima_Suppression:

suppressing non-maximum points

Ojard_Watershed:

watershed analysis per E. Ojard; has hierarchical segmentation

Sobel Operator:

Sobel edge operator

Manipulation

The ``Manipulation'' category features functions for various image manipulations. Many functions involve pointwise operations on one or two volumes, such as thresholding, zeroing out low-valued voxels, taking the pointwise product of two different volumes, etc. Others combine two different volumes, such as volume algebra. Functions for gray-scale interpolation, shrinking/expanding images, adding noise, et al., appear in this category.

The list of functions here is:

Copy_volume:

copy one volume into another

Volume_algebra:

pointwise functions combining 2 volumes

Thresholding:

intensity thresholding

Complement:

complement a volume

Zero_hi_voxels:

zero out bright voxels

Zero_lo_voxels:

zero out dim voxels

Umbra:

compute the umbra of a volume

Gaussian_Noise_Generator:

add Gaussian noise to a volume

2D_Shading:

make a shaded relief-map of a 2D gray-scale image

Linear_Interpolation:

several 3D linear gray-scale image-interpolation techniques

Demask_Region:

Using a labeled segmented image, form an image containing only one selected region

Nonlinear_Interpolation_weh:

several 3D nonlinear gray-scale image-interpolation techniques

Grayscale Inversion:

invert the gray scale of an image

3D Expand:

form an expanded (zoomed) version of an image

3D Shrink:

form a shrunken (subsampled) version of an image

Save_Largest_Region:

save the largest connected region in an image

Measurement

The ``Measurement'' category features functions for making measurements on regions in images, for making comparisons between two regions, for getting the gray-scale characteristics (histogram) of images, etc.

The list of functions here is:

Error_func:

make various comparisons between two images

Region_properties:

compute gray-scale and shape-based properties of a selected image region

Two_vol_Comparison:

Same as Error_Func? NOT AVAILABLE?

Volume_properties:

Compute various simple properties of a volume

Histogram:

compute various types of histograms of an image

3D_Histogram:

Compute a histogram profile image of all slices in a 3D image

Minimum_bounding_cuboid:

find the bounding box of a region

Branch_Measurements:

Construct a measurement profile of a branching (skeletal) structure

Turnkey

The category ``Turnkey'' contains functions that consist of integrated packages of simpler functions; a typical function in this category uses several simpler functions from the other categories. Most of these functions perform some sort of image segmentation operation.

The list of functions here is:

Region_separation:

separate connected regions

Region_isolation:

extract one large bright region

Concave_gap_filling:

fill concave gaps in regions

LV_extract:

extract left-ventricular chambers

Solid_Blob_Extraction:

extract bright solid blobs

Solid_Blob_Test:

run tests on Solid_Blob_Extraction

3D_Angiogram_Analysis:

extract the 3D coronary arteries

Aortic_root_Removal:

MISNOMER! get a large protruding part of an object (aortic root)

Input/Output

The category ``Input/Output'' contains functions for simple input and output.

The list of functions here is:

Display:

Bring up a display of an image

Load:

Load an image from disk

Save:

Save an image to disk

Example_Category

The category ``Example_Category'' gives an example of an IMPROMPTU functions. Mike Hansen's manual on adding functions to IMPROMPTU refers to this example; in the directory

 /work/psu-sources/manual/impromptu

The list of functions here is:

Example:

a dummy example of an impromptu function

• Filter
  • ``Lowpass'' -- lowpass filter
  • ``Median'' -- a median filter
  • ``maximum homogeneity'' -- the maximum-homogeneity filter
  • ``Symmetric_Nearest_Neighbor'' -- symmetric nearest-neighbor mean filter
  • ``Sigma'' -- sigma filter
  • ``Average_Maximum'' -- average/maximum filter
  • ``Average_Minimum'' -- average/minimum filter
  • ``Salt_and_Pepper_Removal'' -- remove salt and pepper (spike) noise
  • ``Anisotropic_Diffusion'' -- anisotropic diffusion filter
  • ``4D_Filtering'' -- 4D mean filter and grayscale inversion
• Morphology
  • ``Binary_Erosion'' -- binary erosion
  • ``Binary_Dilation'' -- binary dilation
  • ``Binary_Opening'' -- morphological binary opening
  • ``Binary_Closing'' -- morphological binary closing
  • ``Minimum'' -- local minimum
  • ``Maximum'' -- local maximum
  • ``Grayscale_Erosion'' -- grayscale erosion
  • ``Grayscale_Dilation'' -- gray-scale dilation
  • ``Grayscale_Opening'' -- morphological grayscale opening
  • ``Grayscale_Closing'' -- morphological grayscale closing
  • ``Ultimate_Erosion'' -- ultimate erosion
  • ``Conditional_Dilation'' -- conditional dilation
  • ``4D_Morphology'' -- 4D mathematical morphology functions
  • ``tophat transform'' -- detect peaks
• Topology
  • ``3D_conn_comp'' -- three-dim. connected components analysis
  • ``2D_conn_comp'' -- two-dim. connected components analysis
  • ``Cavity_Deletion'' -- delete interior cavities
  • ``Homotopic_Thinning'' -- thin an input volume
  • ``Homotopic_Thickening'' -- conditional homotopic thickening
• Segmentation
  • ``3D_Adaptive_thresholding'' -- automatically threshold a 3-D volume
  • ``Hysteresis_Threshold'' -- thresholding using hysteresis
  • ``Region_growing'' -- segment a volume using region growing
  • ``Relaxation_Labeling'' -- segment an image using relaxation labeling
  • ``nonmax suppression'' -- non-maximum suppression
  • ``Kirsch edge operator'' -- Kirsch edge operator
  • ``Sobel edge operator'' -- Sobel edge operator
  • ``diff. of Gaussians'' -- difference of Gaussians
  • ``border of binary vol'' -- find borders in a binary volume
  • ``morpholog. gradient'' -- morphological gradient
• Manipulation
  • ``copy vol1 to vol2'' -- copy one volume into another
  • ``two-volume ops'' -- pointwise functions combining 2 volumes
  • ``threshold'' -- intensity thresholding
  • ``Gray-scale inversion'' -- Gray-scale inversion
  • ``histogram'' -- compute a histogram of a volume
  • ``3-D histogram'' -- compute a 3-D histogram
  • ``umbra'' -- compute the umbra of a volume
  • ``zero hi voxels'' -- zero out bright voxels
  • ``zero lo voxels'' -- zero out dim voxels
  • ``complement'' -- complement a volume
  • ``min.-bounding cuboid'' -- finding object bounds
  • ``add Gaussian noise'' -- add Gaussian noise to a volume
  • ``volume properties'' -- Compute various properties of a volume
  • ``3D_Shrink'' -- generate a shrunken (subsampled) version of an image
  • ``3D_Expand'' -- generate an expanded (zoomed) version of an image
  • ``Flip'' -- Flip a 3D Image
• Measurement
  • ``two-vol. comparison'' -- find differences between two volumes
  • ``volume statistics'' -- find properties of a segmented volume
  • ``demask region'' -- identifies segmented regions within a volume
  • ``save largest region'' -- extract largest region from segmentation volume
• Turnkey
  • ``LV Extract'' -- extract left ventricular chambers
  • ``solid-blob extract'' -- extract bright solid blobs
  • ``solid-blob test'' -- run tests on solid-blob extract
  • ``object separation'' -- separate connected objects
  • ``region isolation'' -- extract a large bright region
  • ``fill concave gaps'' -- fill concave gaps in objects
• Input/Output
  • ``copy vol1 to vol2'' -- copy one volume into another
• Example_Category

Updated: Fri May 19 14:39:07 CDT