Transcript Noise Reduction In MR Images

CSE
5810
Research Topics in Computer Science and
Medical Imaging
Gordon Devoe - 2014
DEVOE-1
CSE
5810
A Nonlocal Maximum Likelihood
Estimation Method for Rician Noise
Reduction in MR Images
Lili He & Ian Greenshields
February 2009
DEVOE-2
Magnetic Resonance (MR) Images

CSE
5810



MRI is the newest most versatile medical imaging
technique available (very minor fluctuations in
chemical composition can be determined).
High resolution images of a patient’s interior body
without surgery.
Strong magnets and pulses of radio waves used to
manipulate the natural magnetic properties in the
body.
Especially useful for capturing images of the brain,
spine, organs and soft tissue.
DEVOE-3
Magnetic Resonance (MR) Images
CSE
5810
DEVOE-4
Noise in Medical Resonance Images
CSE
5810


Sources
 Thermal noise from the conductivity of the
system’s hardware.
 Inductive losses from the conductivity of the
object being images.
Factor Trade-offs
 Resolution
 Signal-to-noise ratio (SNR)
 Acquisition speed
DEVOE-5
Denoising MR Images
CSE
5810
By denoising magnetic resonance (MR) images we can
improve medical processes including:
 Clinical Diagnosis
 the determination of the nature of a disease
 distinguishing one disease from another.
 Tissue Classification
 Systematic arrangement of similar tissues on the
basis of certain differing characteristics.
 Image Segmentation
 Coming Soon…
DEVOE-6
Research Topics In Medical Imaging

CSE
5810

Image Registration
Image Segmentation
DEVOE-7
CSE
5810
Image Registration Methods: A Survey
Zitová, Barbara and Flusser, Jan. Image and
Vision Computing, ISSN 0262-8856, 2003,
Volume 21, Issue 11, pp. 977 - 1000
DEVOE-8
Image Registration

CSE
5810


Image registration is one of the most fundamental
concepts when using computer science techniques to
compare images with one another.
Image registration is the process of overlaying two or
more pictures/images of the same scene and
comparing them from different viewpoints, times
and/or by different sensors.
The intersection of these images involves a set of
changes influenced by different imaging conditions or
various data sources such as image fusion, change
detection and multi-channel image restoration.
DEVOE-9
Image Registration - Applications

CSE
5810








Multispectral Classification
Environmental Monitoring
Change Detection
Image Mosaicking
Weather Forecasting
Creating super-resolution images
Geographic information systems (GIS)
Cartography
Computer Vision
DEVOE-10
Image Registration – Applications Medical

CSE
5810



Comparing nuclear magnetic resonance spectra data
(NRM) with computer tomography (CT) data.
Monitoring tumor growth
Comparing patient data with anatomical atlases
(digital libraries of anatomy information).
Brain mapping.
DEVOE-11
Image Registration – 3 Types

CSE
5810

Multi-Temporal Image Registration
 Images captured at different times/conditions
 Evaluate changes in the scene
 Tumor evolution monitoring
 Healing therapy
Multi-Modal Image Registration
 Images from same scene are captures with sensors
with different characteristics.
 Integrate the information obtained from different
source streams to gain more detailed scene
representation.
 Combine MRI, ultrasound or CT, magnetic
resonance spectroscopy for radiotherapy & nuclear
medicine applications.
DEVOE-12
Image Registration – 3 Types

CSE
5810
Scene-to-model Image Registration
 Images of a scene and models are compared.
 Models are computer generated and images are
overlaid on the computer model.
 Comparing a patient’s image with a digital model.
DEVOE-13
Image Registration - Steps

CSE
5810



Feature Detection
 Salient and distinctive objects (close-boundary
regions, edges, contours, edges, lines, corners, etc.)
 Features are then represented by their controls
points.
Feature Matching
 Correspondences between the features detected in
the sensed and reference image are established.
Transform Model Estimation
 Type and parameters of mapping-functions
aligning the sensed/reference image are
established.
Image Resampling and transformation
 Sensed image is transformed based on the
established mapping functions
DEVOE-14
Image Registration - Steps
CSE
5810
Feature Detection
Feature Matching
Transform Model
Estimation
DEVOE-15
Image Registration – Feature Detection

CSE
5810
Features
 Salient structures
 Regions (forests, lakes, fields)
 Lines (region boundaries, coastlines, roads, rivers)
 Points (region corners, lines intersections, points on
curves)
Stable in time and stay in fixed position
 Distinct and spread all over the image.
Features - Medical Domain
 Line detection to compare anatomical structures.
 Interactive selection.
 Introduction of extrinsic features (screw markers,
dental adapters, etc.)


DEVOE-16
Image Registration – Feature Matching

CSE
5810

Sensed and references images can be compared using
 Intensity values
 Feature Spatial Distributions
 Feature Symbolic Descriptions
Feature Mapping can be divided in to two main
categories:
 Area-based Feature Detection
 Feature Based Feature Detection
DEVOE-17
Image Registration – Area Based

CSE
5810
Correlation “Template Matching”
 Image intensities are directly matched without any
structural analysis.
 Sensitive to noise, varying illumination, and/or by
using different sensor types.
 Useful for real-time applications and easy to
implement hardware.
figure, imshowpair(peppers(:,:,1),recovered_onion,'blend')
DEVOE-18
Image Registration – Area Based

CSE
5810
Mutual Information
 Statistically dependency between two data sets.
 Comparing anatomical and sensed images of
patient’s body.
MI criterion (bottom)
computed in the
neighborhood of point P.
Maximum of MI shows
correct matching position
(point A). Point B indicates
the false matching position
selected by human
operator.
DEVOE-19
Image Registration – Transform Model Est

CSE
5810




Once features have been discovered in the
sensed/referenced images a mapping function can be
constructed to overlay the sensed image onto the
reference image.
The mapping function is designed in a way to ensure
the correspondence of the control points from the
sensed/reference images are as close as possible.
Global mapping models use all control points for
estimating one set of mapping function parameters for
the entire image.
Local mapping models break the images in to sections
making the function parameters depend on the location
of their support in the image and the mapping function
is defined for each section separately.
Changes in medical images tend to appear locally.
DEVOE-20
Image Registration – Transform Model Est
CSE
5810




Similarity Transform (top left)
Affine Transform (top right)
Perspective Projection (bottom left)
Elastic Transform (bottom right)
DEVOE-21
CSE
5810
Current Methods in Medical Image
Segmentation
Dzung L. Pham, Chenyang Xu, Jerry L. Prince
John Hopkins University
Department of Electrical & Computer
Engineering
DEVOE-22
Image Segmentation
What is image segmentation?
CSE
5810



Images are divided up in to smaller pieces or groups of
pixels.
Segments are used to simplify or change the original
image to a format easier to analyze.
Used to identify objects, lines, curves, etc. within an
image.
Non-Trivial problem - noise, missing data, overlap of
features, etc.
DEVOE-23
Image Segmentation - Medical
CSE
5810
In the medical domain image segmentation is extremely
important when using computers to facilitate the growing
number of images in biomedical-imaging applications.






Quantify tissue volumes.
Reconstruction of soft tissues such as the cerebral
cortex.
Study of anatomical structures.
Treatment planning
Computer-integrated surgery
Growing number of use cases…
DEVOE-24
Image Segmentation - Methods
CSE
5810
Image segmentation can be roughly divided in to eight
categories:







Thresholding Approaches
Region Growing Approaches
Classifiers
Clustering Approaches
Markov Random Fields Models (extension)
Artificial Neural Networks
Atlas-Guided Approaches
DEVOE-25
Image Segmentation - Thresholding
CSE
5810
Image segmentation can be achieved by assigning an
intensity value to every pixel of an image. Pixels can then
be categorized based on whether or not they meet this
threshold value.
DEVOE-26
Image Segmentation - Thresholding

CSE
5810

Medical Uses
 Digital Mammography
Limitations
 Only two classes are generated
(cancerous/healthy).
 Cannot be applied to multichannel images.
 Doesn’t take in to account spatial characteristics.
 Sensitive to noise and intensity inhomogeneities
DEVOE-27
Image Segmentation – Region Growing
CSE
5810
A ‘seed point’ is used to extract pixels from a particular
part of the image and then transformations are made to the
seed to perform further analysis.
 Separate regions that have the same properties.
 Provide original images with good segmentation
results.
 Choose multiple criteria at the same time.
DEVOE-28
Image Segmentation – Region Growing
CSE
5810
Original Figure
155 ~ 255
190 ~ 255
255 ~ 255
DEVOE-29
Image Segmentation - Clustering
CSE
5810
Image segmentation using clustering algorithms is very
similar to segmentation classification methods except
clustering algorithms do not require any training data.
The K-means algorithm iteratively computes a mean intensity for each class and
classifies each pixel of the image to the class with the closest mean.
DEVOE-30
Image Segmentation - Clustering

CSE
5810


Advantages
 Do not require training data (self-learning).
 Lack of spatial modeling provides fast
computation.
Disadvantages
 Sensitive to noise and intensity inhomogeneities.
 Do not directly support spatial modeling.
Potential Algorithms
 K-means Algorithm
 Fuzzy c-means Algorithm
 Expectation-maximization Algorithm
DEVOE-31
Image Segmentation – MRF Models

CSE
5810

Statistical model that can be used within segmentation
methods such as clustering.
Models spatial interaction between nearby pixels.
K-means Segmentation
With MRF
DEVOE-32
Image Segmentation – MRF Models
CSE
5810
Medical images are a good candidate for Markov Random
Field (MRF) models because pixels are generally in the
same area as others from the same class and are widely
used in digital mammograms.
•
•
•
•
A depends on B and D
B depends on D and A
C Depends on E
Etc.
DEVOE-33
Image Segmentation – Deformable Models

CSE
5810

Deformable models delineate region boundaries by
using closed parametric curves or surfaces.
These curves and surfaces deform when under the
influence of internal or external forces
DEVOE-34
Image Segmentation – Deformable Models
Advantages
CSE
5810


Ability to directly generate closed parametric curves
or surfaces from images.
Incorporation of a smoothness constraint that provides
robustness to noise and spurious edges.
Disadvantages


Manual interaction is required to place initial model
and choose appropriate parameters.
Deformable models exhibit poor convergence to
concave boundaries.
DEVOE-35
Image Segmentation – Deformable Models
CSE
5810
Deformable surface used for
the reconstruction of the
cerebral cortex.
Intersection between
deformable surface and
orthogonal slices of the MR
image.
DEVOE-36
Image Segmentation – Atlas Guided

CSE
5810



When an atlas or template is available atlas-guided
image segmentation methods can be employed to
anatomically separate features.
A template can simply be a collection of information
about the anatomy of an image being segmented.
A Template can continually be re-used to segment
other images of the same type.
By using an original image and the pre-segmented
template a ‘warped’ image can be generated by using
linear and non-linear transformations.
DEVOE-37
Image Segmentation – Atlas Guided
CSE
5810



Template Image
Target Image
Warped Template
DEVOE-38
Image Segmentation – Atlas Guided
CSE
5810
Three slices from a MR brain volume image overlaid with
the atlas.
DEVOE-39
Image Segmentation – Future Work

CSE
5810
Future work
 Improving accuracy.
 Incorporating prior information from atlases.

Increasing precision tolerances.
 Combine discrete & continuous spatial-domain
segmentation methods.

Improve computational speed of segmentation
methods.
 Multiscale processing
 Parallelizable methods

Reduce amount of manual interaction.
DEVOE-40
Questions
CSE
5810
DEVOE-41
References
Medical Image Registration
CSE
5810
Image registration methods: a survey by Zitová, Barbara and Flusser, Jan. Image and Vision Computing, ISSN 02628856, 2003, Volume 21, Issue 11, pp. 977 – 1000
Q. Chen, M. Defrise, F. Deconinck, Symmetric phase-only matched filtering of Fourier–Mellin transform for image
registration and recognition, IEEE Transactions on Pattern Analysis and Machine Intellingence 16 (1994) 1156–
1168.
Multimodality image registration by maximization of mutual information by Maes, F; Collignon, A; Vandermeulen,
D; Marchal, G; Suetens, P. IEEE transactions on medical imaging, ISSN 0278-0062, 04/1997, Volume 16, Issue 2,
pp. 187 – 198
Medical Image Segmentation
Current methods in medical image segmentation by Pham, D L; Xu, C; Prince, J L Annual review of biomedical
engineering, ISSN 1523-9829, 2000, Volume 2, p. 315
http://en.wikipedia.org/wiki/Image_segmentation
http://en.wikipedia.org/wiki/Region_growing
http://en.wikipedia.org/wiki/Markov_random_field
Magnetic Resonance Imaging (MRI) Noise Reduction
A nonlocal maximum likelihood estimation method for Rician noise reduction in MR images by He, Lili and
Greenshields, Ian R IEEE transactions on medical imaging, ISSN 0278-0062, 02/2009, Volume 28, Issue 2, pp. 165 172
DEVOE-42