¡@
ILIB2 example program
The image analysis library extends ILIB image
processing to analysis. It emphasises the central role of feature extraction in
practical applications. ILIB2 contains extensive image analysis software
including:
- Advanced Image Utilities
- Blob and Particle Analysis
- Hough Transforms
- Moments
- Texture Operators
- Advanced Morphology
- Fourier Transforms
- Image Segmentation
- Shape functions
- Walsh Transforms
¡@
The image analysis library meets the needs of the advanced developer who requires professional standard tools for developing applications. It has the advantage that the system designer can incorporate advanced image evaluation techniques, such as neural networks, directly.
If your needs go beyond basic image processing into real world applications, ILIB2 provides the routines for advanced image analysis. Compared to ILIB2, much alternative imaging software looks limited. ILIB2 bridges the gap between basic image processing software and the neural imaging and pattern recognition methods in ILIB3.
The ILIB2 library is for advanced analysis of images. In a typical application,
ILIB1 functions would be used for manipulating and enhancing an image before
analysis. Then ILIB2 functions could be used to segment the image and find the
required objects, measuring features such as the number, area, length and shape
of the objects in the image. ILIB2 adds several hundred functions to the image
processing capabilities of ILIB1.
The Operators
Many ILIB2 procedures can be used to create an output image or save memory by
operating in-place on the original image. Morphological and related operators
can iterate for a specified number of times or to their limit. The masked
functions have a pixel by pixel defining overlay to operate on regions of any
shape. Image utilities are common to all the libraries enabling ILIB2 to be used
by itself or in combination with ILIB1 (image processing) and ILIB3 (neural
imaging). All the ILIB image types are supported in ILIB2 including 8, 16 and 32
bit mono, 24 bit true colour and 32 bit floating point. The transforms generally
operate on floating point images or complex floating point image pairs.
The Image Analysis Library includes:
Advanced Morphological Functions:
Bridge Operators: 8, 4 and n connected
bridge; corner bridge; horizontal and vertical bridge; diagonal bridge; NESW and
NWSE bridge; U bridge.
Fatten Operators:
for n iterations.
Shrink Operators:
for n iterations and to limit.
Skeletonization:
for n iterations and to limit.
Thicken: two forms
for n iterations and to limit.
Thinning: standard,
background and exothin; two forms, for n iterations and to limit.
Utility Functions:
Vector mean; vector min and max; vector moment; ROI area; vector standard
deviation; image fold; bilinear interpolation; box ROI; polynomial
interpolation; Hough transform limit.
Blob and Particle Functions:
Count objects and find centres; count objects and loop holes; area of objects;
area and perimeters; perimeters; 4, 8 and average euler numbers; list centre x,y
positions of objects.
Boundary Functions:
Five perimeter finding functions in place or standard.
Colour Functions:
Colour to mono; mono to pseudo-colour; get and set colour mask; copy colour
primary; convert from RGB to other colour systems, e.g. HSI etc.; convert from
other colour system to RGB.
Erosion and Dilation:
Offset generalised dilation; 3x3, 5x5, NxM generalised dilation; NxM extended
erosion and dilation; offset generalised erosion; erode 3x3, 5x5 and NxM;
generalised grey level/colour erosion and dilation; generalised open and
generalised close.
Image Utilities:
Heap and stack check; log image; get and set ROI; set and get image modifier;
binarize image; set image to constant; extract region of interest; zero image or
region; vector and matrix validation; check region limits; test image structure.
Limits and Segmentation:
Column and row pixel sums; masked column and row pixel sums; find pixel ON
block; find top or bottom pixel ON rows (stepped and standard); find left or
right pixel ON columns (stepped and standard); find pixel block; find top or
bottom pixel rows (stepped and standard); find left or right pixel columns
(stepped and standard).
Masked Mathematical and Statistical Functions:
Nth moment about pixel mean; Nth moment from value; masked entropy pixels;
masked entropy region; histogram; histogram over range; kurtosis; arithmetic,
geometric and harmonic means; median; mode; ON pixel count; pixel maximum and
minimum; pixel range; skew; standard deviation; pixel sum; variance.
Memory Utilities:
Range of vector and matrix utilities, as ILIB1.
Moments:
Zeroth moment; x or y moments; x or y Nth moments; arbitrary x and y Nth by Mth
moments; x or y centroid; the seven Hu moments.
Object Shape Features:
Angular shape distribution of object; centre of object; length and width of
object; circularity; eccentricity.
Texture Analysis:
Nine Chebyshev filters and nine Laws filters
Transforms:
2D cosine transform and inverse; 2D sine transform and inverse; 2D fft and
inverse; 2D fast Walsh transform and inverse; Hough transforms; vector
convolution; vector deconvolution; vector correlation; 1D cosine transform and
inverse; 1D sine transform and inverse; 1D fft and inverse; 1D real transform
and inverse; 1D fast Walsh transform and inverse.
Transform Filters:
Bartlett x or y filters; Blackman x or y filters; arbitrary x or y filters;
Hamming x or y filters; Hanning x or y filters; Kaiser x or y filters;
rectangular x or y filters.
