计算机视觉与医学图像处理
Compare ei and ej Cluster analysis to group similar expression patterns Define novel genes that are similar to known genes Characterize cluster functions Find genes which fit a prototypical pattern over samples
Outline Extraction
original
x f K f ( x 1, y ) f ( x, y )
y f K f ( x, y 1) f ( x, y )
MAX max( x f , y f )
Boundary Extraction
Extraction of Regions of Interest (ROI)
Skeletonization
Structure Representation and Analysis
Structure Representation Structure Organization Quantitative Computation
Image Processing – Enhancement
Control1
73.0MB
95
291
5
<1
464
Control2
80.2MB
104
121
6
<1
309
Control3 114.4MB
147
6
8
<1
269
Humliv
24.4MB
56
72
4
<1
179
CVGIP2002, Aug. 25-27, 2002
Bioinformatics
Use information techniques to solve biological problem Reasons to exist
Microarray Data
s1 e1 e11 e2 e21 e3 e31 e4 e41
s2 e12 e22 e32 e42
s3 e13 e23 e33 e43
s4 e14 e24 e34 e44
Organized as an m x n matrix M m: no. of genes; n: no. of samples Mij: expression level of gene i in sample j Row ei: expression pattern of gene i Column sj: expression pattern of sample j Ratioing: eij = log (Mij/gi)
gi: expression level in a control The logarithm (log) is useful for normalization of expression profile
Discovering Microarray Data (I)
Genes w/ similar expression pattern over all samples
生命的介面
電腦視覺與醫學影像
OUTLINE
Learn from Living Mechanisms
Computer Vision: Modeling Human Vision Human Vision vs. Computer Vision Image processing: Artificial pattern recognition Medical Imaging: Mining physiological info
Human Vision System
lens: 水晶體 iris: 虹膜 cornea: 角膜 retina: 視網膜 optic nerve: 視覺神經 cone: photopic (bright-light) vision rod: scotopic (dim-light) vision
Synthetic Camera Model
Image Processing (I)
Fundamental units
transverse plane
2D – pixel 3D – voxel
coronal plane sagittal plane
Orthogonal views: transverse (or axial), coronal, sagittal Image processing: preprocessing, segmentation, object representation and recognition, quantitative analysis
Image Processing (II) – Slices vs. Projections
Maximum Intensity Projection (MIP)
Y
Z
X
Weighted-Sum Projection
Image Processing Procedure
Visualization Tools
LCA (Left Circumflex Arteries) Analysis
Original lca_40
Segmented lca_40
Segmentation
Thinned lca_40
Surface rendered lca_40
MR-PET Fusion
Pipelined Data-based Visualization
Procedural Flow of Image Processing
Segmentation Representation and description
Preprocessing Result Problem Domain Knowledge base Recognition and interpretation
Computer Vision
Thinking Image Processing Pattern Recognition
Visual Perception
Phenomena of Perception
Digital Images
Medical Display Workstations
Dual-Monitor Display Workstation
Medical Imaging
Objective: assist the physician in study/identification of anomaly in the organism Medicine domain knowledge + image processing + visualization Topological & geometrical analysis Validation issues – how to obtain ground truth? Need of human interaction Scanning techniques – CT, MR, PET, Ultrasound Resolution issue: delta z usually significantly larger than delta x and delta y
Image Acquisition
3-D Medical Image Processing & Analysis
Extraction of Regions of Interest
Enhancement/ Filtering
Segmentation
Post Processing Extracted Structures
Discovering Microarray Data (II)
Genes w/ unusual expression levels in a sample
Examine the change of expression levels in certain si vector Determine outlier genes of interest
monitor pre- and post-state of genes Determine possible clusters of samples Discover genetic clues that may lead to find subtype of a cancer Classify or diagnose cancers by relating them to information from expression patterns of known cancerous and non-cancerous tissues
3D CT Images and Analysis
Human liVGIP2002, Aug. 25-27, 2002
Results: 3D CT Image Analysis
Image Name
Image Size
Time usage (in seconds) on PC (CPU:400Mhz) Sig Seg Cav 8 4 8 4 9 4 4 4 64 20 69 18 98 24 42 12 Thin Rep Man Total
A great deal amount of biological data Identification of known information Discovery of hidden pattern Quantification