MedIX

1
MedIX Summer 06

• Lucia Dettori (room 745)
• ldettori_at_cti.depaul.edu

2
Projects

• Texture classification
• What has been done
• Things I would like to explore next
• Connection to other projects
• Evaluations of segmentation algorithms

3
Done so far

• Given a pre-segmented organ region, can you tell
  me what it is kidney, heart etc?
• It depends on its texture
• Identify image features that give texture
  information
• Find rules that distinguish the texture features
  of one organ from another

4
Texture Classification Process at a glance
Organ/Tissue Segmented Image
5
Step1 Segmentation and cropping
The image might need to be cropped, when using
filters that are sensitive to areas of high
contrast (background)
Active Contour Mapping (Snakes) a boundary
based segmentation algorithm
Organs Backbone Heart Liver Kidney Spleen
Segmented 140 50 56 55 39
Cropped 363 446 506 411 364
6
Step2 Filtering the image

• For example
• Co-occurrence matrices
• Run-length matrices
• Wavelets
• Ridgelets
• Curvelets

Organ/Tissue Segmented Image
Wavelet transform
Averages Horizontal Activity
Vertical Activity Diagonal Activity
7
Haar Wavelet
A D


A D
Original image
9 7 3 5
A
D
A
D
8 4
1 -1
Averages
Details
6
2
1 -1
6 2 1 -1
Wavelet coefficients
AA AD
DA DD
8
Step3 Texture features extraction
Organ/Tissue Segmented Image

• For example
• Mean, standard deviation, energy, entropy etc..

Array of texture descriptors T1, T2, T3, ,
Tn
9
Step4 - Classification
Organ/Tissue Segmented Image
The process of identifying a region as part of a
class (organ) based on its texture properties.
Predicts the organ from the values of the texture
descriptors Training / Testing
10
Step5 Evaluating the classifier
Misclassification matrix
Actual Category Backbone Heart Liver K
idney Spleen Total Predicted Backbone 182
6 1 6 0 195 Category Heart
3 18 4 0 0 25 Liver 0
3 30 1 7 41 Kidney 10
4 0 49 1 64 Spleen 0
0 4 0 8 12 Total 195 31
39 56 16 337
Performance Measures
11
Organ Descriptor Sensitivity Specificity Precision Accuracy
Backbone Wavelet 82.6 96.1 82.6 93.7
Backbone Ridgelet 91.5 99.3 96.8 98.0
Backbone Curvelet 99.4 98.8 95.3 98.9
Heart Wavelet 59.0 92.1 67.0 85.0
Heart Ridgelet 82.5 97.5 88.5 94.6
Heart Curvelet 89.7 99.0 95.5 97.1
Kidney Wavelet 77.7 91.4 69.9 88.6
Kidney Ridgelet 95.4 93.3 82.0 93.8
Kidney Curvelet 96.0 98.1 93.5 97.6
Liver Wavelet 87.3 94.4 82.6 92.8
Liver Ridgelet 86.9 95.9 84.4 94.0
Liver Curvelet 95.9 98.5 94.3 98.0
Spleen Wavelet 65.5 94.3 69.7 89.5
Spleen Ridgelet 76.9 97.6 88.0 93.8
Spleen Curvelet 91.8 98.9 94.9 97.6
Average Wavelet 74.4 93.7 74.4 89.9
Average Ridgelet 86.6 96.7 88.0 94.8
Average Curvelet 94.6 98.7 94.7 97.9
12
Things I would like to explore
Different modalities
Organ/Tissue Segmented Image
Different patients Different organs Abnormal
texture
Gabor filters Fractal Dimensions
13
Connections to other project

• Can we use wavelet, ridgelet, curvelet-based
  texture descriptors for content based image
  retrieval?
• Can we use these descriptors in the volumetric
  segmentation?
• Instead of many 2D images, can we use the same
  process for 3D stack of slices?

14
Projects

• Texture classification
• Evaluations of segmentation algorithms
• What has been done
• Things I would like to explore next
• Connection to other projects

15
Texture segmentation

• Given an image, can you tell me how many organs
  you have?
• That was easy enough. Can youtell which organs
  they are?
• Identifying regions with similar texture
• Identifying which texture it is to label the organ

16
A couple of key questions

• Can you do it better by varying a parameter? How
  do you choose the values of your segmentation
  parameters?
• If it looks better is it really better?

17
A couple of key questions

• Parameter optimization
• Performance evaluation

18
3
1
0.56
0.87
4
2
0.50
0.75
19
How do I decide what the optimal value of the
parameter is? How good a segmentation is it?
GroundTruth
Regionskey
Machine Segmentations
Increasing value of a segmentation parameter
20
The goodness metric

• A single value that assigns a rating to a
  particular segmentation based on how well the
  machine segmented regions match the regions in
  the ground truth images

21
Region Categories

• Ground Truth vs. Machine Segmented
• Correctly Detected
• Over Segmented
• Under Segmented
• Missed
• Noise

GT
MS
22
OVER SEGMENTED
CORRECTLY DETECTED
Index for each region
UNDER SEGMENTED

• A Missed region is a GT region that does not
  participate in any instance of CD, OS, or US
• A Noise region is an MS region that does not
  participate in any instance of CD, OS, or US

23
The Goodness Metric

• good Correct Detection Index
• bad 1-Correct Detection Index
• goodness good-badweight

1.0
Ceiling CDind
Weight Range CDind-1
Floor 2CDind-1
-1.0
24
How can we use the metric?

• Create a set of ground truth mosaic using
  radiologist-labels images of pure patches of
  organ tissues
• Apply segmentation algorithm
• Optimize the segmentation parameters using the
  metric
• Apply optimized algorithm to the real image

25
Region key
Ground Truth
T3000 GM .73
T1000 GM - .94
T2000 GM - .02
T4000 GM .74
T5000 GM .75
T6000 GM .08
26
Done so far

• Used the metric on a block-wise walevet-based
  segmentation algorithm on some sample mosaic

27
To be done

• Fully test the metric on a wide range of
  segmentation algorithms
• Decouple the various components of the metric and
  test the individual performance measures instead
  of the overall score
• Extend the metric to measure one region vs
  background segmentation

28
To be done

• Improve the wavelet-based algorithms we have
  implement to include other texture features
• Explore and compare other texture-based
  segmentation algorithm
• Use regions and metric to calculate changes in
  time of an abnormal region

29
Connections to other projects

• Use one of these algorithms to create a rough
  segmentation that will generate the starting
  point for a more sophisticated segmentation
  algorithm.

30
Some references

• Wavelet-based Texture Classification of Tissues
  in Computed Tomography, L. Semler, L, Dettori,
  and Jacob Furst.18th IEEE International
  Symposium on Computer-based Medical Systems,
  Dublin, Ireland, June 2005.
• Ridgelet-based Texture Classification in
  Computed Tomography, L. Semler, L. Dettori. and
  W.Kerr. 8th IASTED International Conference on
  Signal and Image Processing, Honolulu, HW, August
  2006.
• Curvelet-based Texture Classification of Tissues
  in Computed Tomography, L. Semler, L. Dettori.
  International Conference on Image Processing,
  Atlanta, GA, October 2006.
• A Comparison of Wavelet-based and Ridgelet-based
  texture classification of Tissues in Computed
  Tomography, with Lindsay Semler, International
  Conference on Computer Vision Theory and
  Applications, Setubal, Portugal, February 2006
• A Methodology and Metric for Quantitative
  Analysis and Parameter Optimization of
  Unsupervised, Multi-Region Image Segmentation,
  William Kerr, Lucia Dettori, and Lindsay Semler,
  8th IASTED International Conference on Signal and
  Image Processing, Honolulu, HW, August 2006.