MEDICAL IMAGE PROCESSING
ANALYSIS OF SHAPE
CHAPTER 5
MEDICAL IMAGE PROCESSING
- Several human organs and biological structures possess
readily identifiable shapes like heart, brain, kidney or
bone.
Normally, don’t deviate much from an “average” shape
- Disease processes can affect the structure of organs
→causing deviation from expected or average shape
- Abnormal entities tend to demonstrate differences in
shape between benign and malignant conditions.
- But biological entities demonstrate wide ranges of
manifestation with significant overlap between their
characteristics for various categories
In 3D-to-2D projection, shapes of objects in
biomedical image will be affected by superimposition of
multiple objects
-2-
5.1. Representation of Shapes and ContoursMEDICAL IMAGE PROCESSING
3
- The most general form of representation of a
contour is by mean of (x, y) coordinates of the
digitized pixels along the contour
+ Not related to gray levels of contour pixels
+ May be depicted as binary of bilevel image
{ ( ), ( )}, 0,1,2..., 1x n y n n N=−
5.1. Representation of Shapes and ContoursMEDICAL IMAGE PROCESSING
-4-
5.1.1. Contour signature (ký số của đường bao)
- Distances from each contour point to a reference point
- Reference point: centroid (trọng tâm) or center of mass
of the contour
- The signature of the contour
11
00
11
( ) ( )
NN
nn
x x n y y n
NN
−−
==
==

22
( ) ( ) ( )d n x n x y n y
= − + −
5.1. Representation of Shapes and ContoursMEDICAL IMAGE PROCESSING
-5-
A contour represented by its boundary points z(n)
and distances d(n) to its centroid