Prostate Edge Detection Using a Knowledge Base

1
Prostate Edge Detection Using a Knowledge Base

• Supervisor
• Prof. Magdy Salama
• By
• Joseph Awad
• November 09

2
Overview

• Introduction
• Algorithm Description
• Contrast Enhancement
• Seed Point Localization
• Knowledge Base Rules
• Morphological Opening
• Boundary Detection by Radial Scanning
• Numerical Results
• Conclusion

3
Introduction

• Prostate cancer is diagnosed every 2.75 minutes,
  approximately 190,000 new cases each year.
• It is the most commonly diagnosed cancer in
  America among men. More than 30,000 American men
  lose their lives to prostate cancer each year,
  one death every twenty minutes.
• Prostate cancer incidence rates increased 192
  between 1973 and 1992.

4
Algorithm Description
5
Contrast Enhancement

• Sticks Technique
• For N N neighbourhood in the image, there are
  2N - 2 short lines that pass through the central
  pixel, with N pixels in length. (3217)

6
Seed Point Localization

• The intensity level of the prostate is low with
  respect to its surrounding area.
• The prostate is not in perimeter of the image but
  not necessarily in the medial.

7
Knowledge Base RulesWhy ?
8
Knowledge Base Rules

• The intensity level of prostate is less than the
  area around it.
• Radial scanning of the image from seed point and
  removes any edge, which represents light to dark
  transitions.

9
Morphological Opening

• There are still some false edges most of them
  have short lengths.
• To remove these false edges, the short linked
  pixels with areas less than 50 pixels can be
  eliminated.

10
Boundary Detection by Radial Scanning

• It might be still some false edges in the edge
  map. It is known that the prostate has a smooth
  curvature shape.
• This algorithm analyses the obtained edge and
  filter out any pixels violate this knowledge. The
  final step in this algorithm is to use these edge
  pixels and interpolate the missing parts and
  found the best spline fit these pixels.

11
Numerical Results Example 1
12
Numerical Results Example 2
13
Numerical Results Example 3
14
Numerical Results Example 4
15
Numerical Results Example 5
16
Conclusion

• The proposed algorithm is built on knowledge
  base.
• It is fully automated.
• It simulates an expert eyes.
• The more knowledge base rules we apply, the
  better results we can get.

17
Thank you
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
(No Transcript)
31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
(No Transcript)
35
(No Transcript)
36
(No Transcript)
37
(No Transcript)
38
(No Transcript)
39
(No Transcript)
40
(No Transcript)
41
(No Transcript)
42
(No Transcript)
43
(No Transcript)
44
(No Transcript)
45
(No Transcript)
46
(No Transcript)
47
(No Transcript)