Transcript (Medical Image Processing Lab.) Chuan

Medical Image Analysis
Chapter 1 Introduction
Chuan-Yu Chang (張傳育)Ph.D.
Dept. of Computer and Communication Engineering
National Yunlin University of Science & Technology
[email protected]
http://mipl.yuntech.edu.tw
Office: ES709
Tel: 05-5342601 Ext. 4337
Medical Image Analysis
Atam P. Dhawan, Ph.D.
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
Introduction
Image Formation
Interaction of Electromagnetic Radiation
with Matter in Medical Imaging
Medical Imaging Modalities
Image Reconstruction
Image Enhancement
Image Segmentation
Image Representation and Analysis
Image Registration
Image Visualization
Current and Future Trends in Medical
Imaging and Image Analysis
Exercises in MATLAB
References
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
Image Databases on Website
2
Imaging in Medical Sciences


Imaging is an essential aspect of medical
sciences for visualization of anatomical
structures and functional or metabolic (新陳代謝)
information of the human body.
Structural and functional imaging of human body
is important for understanding the human body
anatomy, physiological processes, function of
organs, and behavior of whole or a part of organ
under the influence of abnormal physiological
conditions or a disease.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
3
Medical Imaging



Radiological sciences in the last two decades have witnessed a
revolutionary progress in medical imaging and computerized
medical image processing.
Advances in multi-dimensional medical imaging modalities
 X-ray Mammography
 X-ray Computed Tomography (CT)
 Single Photon Computed Tomography (SPECT)
 Positron Emission Tomography (PET)
 Ultrasound
 Magnetic Resonance Imaging (MRI)
 functional Magnetic Resonance Imaging (fMRI)
Important radiological tools in diagnosis and treatment evaluation
and intervention of critical diseases for significant improvement in
health care.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
4
Medical Imaging and Image Analysis



The development of imaging instrumentation has
inspired the evolution of new computerized image
reconstruction, processing and analysis methods for
better understanding and interpretation of medical
images.
The image processing and analysis methods have been
used to help physicians to make important medical
decision through physician-computer interaction.
Recently, intelligent or model-based quantitative image
analysis approaches have been explored for computeraided diagnosis to improve the sensitivity and specificity
of radiological tests involving medical images.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
5
A Multidisciplinary Field



Medical imaging in diagnostic radiology has evolved as a result of
the significant contributions of a number of different disciplines from
basic sciences, engineering, and medicine.
A number of computer vision methods have been developed for a
variety of applications in image processing, segmentation, analysis
and recognition.
However, computerized image reconstruction, processing and
analysis methods have been developed for medical imaging
applications


Require specialized knowledge of a specific medical imaging modality
that is used to acquire images.
The application-domain knowledge has been used in developing
models for accurate analysis and interpretation.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
6
A Multidisciplinary Paradigm
Physiology and Current
Understanding
Physics of Imaging
Applications and
Intervention
Instrumentation
and Image Acquisition
Computer Processing,
Analysis and Modeling
生物醫學影像的智慧分析與判讀，須對影像的取得過程及原理有一定的認識。
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
7
Medical Imaging Modalities


The objective of medical imaging is to acquire useful information about
the physiological processes or organs of the body by using external or
internal sources of energy.
Energy source

Internal




External



Nuclear medicine imaging use an internal energy source through an emission
process to image the human body.
Radioactive pharmaceuticals (放射性藥劑)are injected into the body to interact with
selected body matter or tissue to form an internal source of radioactive energy that
is used for imaging.
Ex. Single Photon Emission Computed Tomography (SPECT), PET
Anatomical imaging are based on the attenuation coefficient(衰減係數) of radiation
passing through the body
Ex. X-ray radiographs and Computed Tomography (CT)
Combination

MRI uses external magnetic energy to stimulate selected atomic nuclei. The
excited nuclei become the internal source of energy to provide electromagnetic
signals for imaging through the process of relaxation.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
8
Electromagnetic Radiation Spectrum
Radio
Waves
103
TV
Waves
102 101
1
Radar
Waves
10-1 10-2
Microwaves
10-3
Infrared
Rays
Visible Ultraviolet
Light
Rays
-6
10-4 10-5 10
10-7
10-8
Gamma
Rays
X-rays
Cosmic
Rays
10-9
10-10 10-11 10-12 10-13 10-14
1017
1018 1019 1020
Wavelength in meters
105 106 107 108
11
109 1010 10
14
1012 1013 10
1015 1016
1021 1022
Frequency in Hz
-8
10-10 10-9 10 10-7 10-6 10-5
10-4
-1
10-3 10-2 10
1
101
102
103
104
105
106
107
Energy in eV
MRI
X-ray
Imaging
Gamma-ray
Imaging
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
9
Medical Imaging Modalities

Radiation/Imaging Source

External





Internal



X-ray Radiography
X-ray CT
Ultrasound
Optical: Reflection, Transillumination
SPECT
PET
Mixed



MRI, fMRI
Optical Fluorescence
Electrical Impedance
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
10
Medical
Imaging
Modalities
Classification of
medical imaging
modalities
Sourceof Energy
Used for Imaging
Internal
External
X-Ray Radiographs
Nuclear Medicine:
Single Photon
EmissionTomography
(SPECT)
X-Ray Mammography
X-Ray Computed
Tomography
Nuclear Medicine:
Positron Emission
Tomography
(PET)
Combination:
External and
Internal
Magnetic Resonance
Imaging: MRI, PMRI,
FMRI
Optical Fluorescence
Imaging
Electrical Impedance
Imaging
Ultrasound Imaging and
Tomography
Optical Transmission
andTransillumination
醫學影像處理實驗室(Medical
Imaging
Image Processing Lab.) Chuan-Yu Chang Ph.D.
11
Medical Imaging Information

Anatomical






X-Ray Radiography
X-Ray CT
MRI
Ultrasound
Optical
Functional/Metabolic






SPECT
PET
fMRI, pMRI
Ultrasound
Optical Fluorescence
Electrical Impedance
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
12
Medical Imaging Modalities (cont.)



There is no perfect imaging modality for all
radiological applications and needs.
Each medical imaging modality is limited by the
corresponding physics of energy interactions with
human body, instrumentation and often physiological
constraints.
The performance of an imaging modality for a
specific test or application is characterized by
sensitivity and specificity factor.


Sensitivity of a medical imaging test is defined primarily by
its ability to detect true information.
The specificity for a test depends on its ability to not detect
the information when it is truly not there.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
13
Medical Imaging: From Physiology to
Information Processing


Understanding Imaging Medium
 The information about imaging medium may involve static or
dynamic properties of the biological tissue.
 Tissue density is a static property, blood flow, perfusion and
cardiac motion are examples of dynamic properties.
Physics of Imaging
 X-ray imaging modality uses transmission of X-rays through the
body as the basis of imaging.
 Single Photon Emission Computed Tomography (SPECT) uses
the emission of gamma rays resulting from the interaction of a
radiopharmaceutical substance with the target tissue.
 The SPECT and PET imaging modalities provide images that are
poor in contrast and anatomical details.
 X-ray CT images modality provides shaper images with highresolution anatomical details.
 The MR imaging modality provides high-resolution anatomical
details with excellent soft-tissue contrast.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
14
Medical Imaging: From Physiology to
Information Processing (Cont.)



Imaging Instrumentation
 Defining the image quality in terms of signal-to-noise ratio,
resolution and ability to show diagnostic information.
 An intelligent image formation and processing technique should
be the one that provides accurate and robust detection of
features of interest without any artifacts to help diagnostic
interpretation.
Data Acquisition Methods for Image Formation
Image Processing and Analysis
 Aimed at enhancement of diagnostic information to improve
manual or computer-assisted interpretation of medical images.
 Interactive and computer-assisted intelligent medical image
analysis methods can provide effective tools to help the
quantitative and qualitative interpretation of medical images for
differential diagnosis, intervention and treatment monitoring.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
15
General Performance Measures

A conditional matrix for defining four basic
performance measures as defined in the text
True Condition
Object is
observed.
Object is
Object is
present.
NOT present.
True
Positive
False
Positive
False
Negative
True
Negative
Observed
Information
Object is
NOT
observed.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
16
General Performance Measures
(cont.)

Positive observation


Negative observation


The object was observed in the test.
The object was not observed in the test.
True condition

The actual truth, whereas an observation is the
outcome of the test.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
17
General Performance Measures
(cont.)

Four basic measures

True Positive Fraction (TPF): TPF=Notp/Ntp


False Negative Fraction (FNF): FNF=Nofn/Ntp


The ratio of the number of negative observations to the
number of positive true-condition cases.
False Positive Fraction (FPF): FPF=Nofp/Ntn


The ratio of the number of positive observations to the number
of positive true-condition cases.
The ratio of the number of positive observations to the number
of negative true-condition cases.
True Negative Fraction (TNF): TNF=Notn/Ntn

The ratio of the number of negative observations to the
number of negative true-condition cases.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
18
General Performance Measures
(cont.)

It should be noted that



Sensitivity


True Positive Fraction, TPF
Specificity


TPF+FNF=1
TNF+FPF=1
True Negative Fraction. TNF
Accuracy=(TPF+TNF)/Ntot
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
19
ROC (Receiver Operating
Characteristic)

ROC (Receiver Operating Characteristic)

A graph between TPF and FPF is called ROC
curve for a specific medical imaging or diagnostic
test for detection of an object.
TPF
a
b
c
TNF
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
20
Example of Performance Measure



Assume that 100 female patients were examined through
the X-ray mammography test.
The X-ray mammography images were observed by a
physician to classify into one of the two classes: normal
and cancer.
The object is to determine the basic performance measures
of the X-ray mammography test for detection of breast
cancer.



Total number of patients=Ntot=100
Total number of patients with biopsy proven cancer (true
condition of object present)=Ntp=10
Total number of patients with biopsy proven normal tissue
(true condition of object NOT present)=Ntn=90
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
21
Example of Performance Measure
(cont.)




Out of the patients with cancer Ntp, the number of
patients diagnosed by the physician as having cancer =
Number of True Positive cases = Notp =8.
Out of the patients with cancer Ntp, the number of
patients diagnosed by the physician as normal =
Number of False Negative cases = Nofn =2.
Out of the normal patients Ntn, the number of patients
rated by the physician as normal = Number of True
Negative cases = Notn =85.
Out of the normal patients Ntn, the number of patients
rated by the physician as having cancer = Number of
False Positive cases = Nofp =5.
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
22
Example of Performance Measure
(cont.)






True Positive Fraction (TPF)= 8/10 = 0.8
False Negative Fraction (FNF)= 2/10 = 0.2
False Positive Fraction (FPF)= 5/90 = 0.0556
True Negative Fraction (TNF)= 85/90 = 0.9444
TPF+FNF=1.0
FPF+TNF=1.0
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
23
Biomedical Image Processing and
Analysis

A general-purpose biomedical image-processing
and image analysis system must have three basic
components:

Image-acquisition system


Digital computer


Usually converts a biomedical signal or radiation carrying the
information of interest to a digital image.
Usually large memory units that are used to store digital
images for further processing.
Image display environment

The output image can be viewed after the required processing,
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
24
Biomedical Image Processing and
Analysis (cont.)

General schematic of biomedical image analysis system
Bio/Medical
Image
Acquisition
System
Digital Computer
And
Image Processing
Unit
Display
Unit
Scanner
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
25
Image Processing Task:
Feature Enhancement
Enhanced image
through feature adaptive
contrast enhancement
algorithm
Enhanced image through
histogram equalization
method
醫學影像處理實驗室(Medical Image Processing Lab.) Chuan-Yu Chang Ph.D.
26