Module -6
Boundary Description & Object Recognition
Topic : Texture
MR. MRUGENDRA VASMATKAR
Assistant Professor, EXTC
VES Institute of Technology
References
1. Milan Sonka ,Vaclav Hlavac, Roger Boyle, “Image Processing, Analysis, and Machine Vision” Cengage Engineering,
3rd Edition, 2013
2. Digital image processing – Ramesh Jayraman
3. Digital Image processing - Sreedhar
Marks in Exam :- 30 Marks – Book 1
Numerical on cooccurance matrix ( book -1)
What is texture
The concept of texture is intuitively obvious but has no precise definition.
Texture is something consisting of mutually related elements.
Texture primitives (or texture elements, texels) are building blocks of a
texture.
Texture description is scale dependent.
Texture can be described by its tone and structure.
Tone ... based on pixel intensity properties.
Structure ... describes spatial relationships of primitives.
Texture can be described by the number and types of primitives and
by their spatial relationships.
Classification of texture
Weak textures
have small spatial interactions between primitives, can be described by
frequencies of primitive types appearing in some neighborhood;
statistical approaches are adequate for weak texture description.
Strong textures
spatial interactions among primitives are somewhat regular;
frequency of occurrence of primitive pairs in some spatial relationship can
be used for description;
frequently accompanied by exact definition of texture primitives.
Description method
1. Definition of a constant texture:
An image region has a constant texture if a set of its local properties in that region is
constant, slowly changing, or approximately periodic.
2. Two basic texture description approaches:
1. Statistical
2. Syntactic
Fine Texture
Coarse Texture
Artificial Textures
Texture tone and structure are not independent; textures always display both tone and structure even
though one or the other usually dominates, and we usually speak about one or the other only.
Statistical Texture description
1. Statistical texture description methods describe texture in a form suitable for
statistical pattern recognition
2. Each texture is described by a feature vector of properties which represents a point in
multi-dimensional feature space
3. The aim is to find a deterministic or probabilistic decision rule assigning a texture to
some specific class
4. One of the simplest approaches for describing texture is to use statistical moments of
the intensity histogram of an image or region.
Co-occurance Matrix
• A cooocurance method of texture description is based on the repeated occurance of
some grey level configuration varies rapidly with distance in fine texture and slowly
in coarse texture
Co-occurance Matrix
Grey Level Image
Algorithm :- Cooccurance method for texture description
Co-occurrence matrices
Edge frequency
1. Comparison of edge frequency can be used for describing the texture
2. Edges can be used as a micro edges using small edge operator masks
3. Or edges can be used as a macro edges using large edge operator masks
4. Robert’s gradient can be used or
5. It can be used as a function of distance between pixels.
6. It can be computed for any subimage f defined in a neighbourhood N for variable distance d:
Algorithm : Edge frequency algorithm
Primitive Length ( Run length)
Primitive Length ( Run length)
Laws' Texture Measures
Developed by Kenneth Ivan Laws at the University of Southern
California in 1979.
Used in many diverse Applications as Texture Energy Measures.
Computed by first applying small convolution kernels to a digital
image, and then performing a nonlinear windowing operation.
The 2-D convolution kernels, used for texture discrimination are
generated from the set of 1-D convolution kernels of length five.
Laws' Texture Measures
L5(Level) = [ 1 4 6 4 1 ] Level Detection
E5(Edge) = [ -1 -2 0 2 1 ] Edge Detection
S5(Spot) = [ -1 0 2 0 -1 ] Spot Detection
R5(Ripple) = [ 1 -4 6 -4 1 ] Ripple Detection
W5(Wave) = [ -1 2 0 -2 1 ] Wave Detection
2-D convolution kernels are generated by convolving a vertical 1-D
kernel with a horizontal 1-D kernel.
May 9, 2025 18
Milan Sonka ,Vaclav Hlavac, Roger Boyle, “Image Processing,
Analysis, and Machine Vision” Cengage Engineering, 3rd
A listing of all 5x5 kernel names is given below.
Edition, 2013
Except L5L5 , others are zero-sum.
L5 L5 E5 L 5 S5 L 5 W5 L5 R 5 L5
L5 E5 E5 E5 S5 E 5 W5 E5 R 5 E5
L 5 S5 E 5 S5 S5 S5 W5 S5 R 5 S5
L5 W5 E5 W5 S5 W 5 W5 W5 R 5 W5
L5R 5 E5 R 5 S5 R 5 W5 R5 R 5R 5
19
May 9, 2025
END
