image segmentation of multiple objects and their compartments jerry l. prince image analysis and...

58
Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL) http://iacl.ece.jhu.edu Johns Hopkins University © 2010

Upload: augustus-watts

Post on 18-Dec-2015

218 views

Category:

Documents


2 download

TRANSCRIPT

Page 1: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Image Segmentation of Multiple Objects and Their Compartments

Jerry L. PrinceImage Analysis and Communications

Laboratory (IACL)http://iacl.ece.jhu.edu

Johns Hopkins University© 2010

Page 2: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Acknowledgments• Chenyang Xu• Dzung Pham• Xiao Han• Duygu Tosun• Bai Ying• Daphne Yu• Kirsten Behnke• Xiaodong Tao• Sarah Ying• Xian Fan

• Susan Resnick• Mike Kraut• Maryam Rettmann• Christos Davatzikos• Nick Bryan• Aaron Carass• Ulisses Braga-Neto• Lotta Ellingsen• Pierre-Louis Bazin

Funding sources: NSF, NIH/NINDS, NIH/NIA

Page 3: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 4: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 5: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Conventional Structural Image

www.medical.philips.com

Page 6: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Segmentation of Brain Structures

Volumetric MR Data

Subcortical Structures

Cortex

TOADS CRUISE

Bazin and Pham, TMI, 2007

Bazin and Pham, MedIA, 2008

Xu et al., TMI, 1999

Han et al., NeuroImage, 2004

Tosun et al., NeuroImage, 2006

Page 7: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Thalamic nuclei

Cerebellar lobules

Multi-Compartment Anatomy

Page 8: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Cell counting

Circuit board inspection

Retinal examination Traffic camera

Satellite imageryAerial photographs

Other Multi-Object Scenarios

Page 9: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 10: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Cortical Surface Segmentation

Page 11: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Partial Inflation

Page 12: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Ventricle Segmentation

Page 13: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

• Parametric deformable models (PDMs)

─explicit parameterization

• Geometric deformable models (GDMs)– implicit

representation

Deformable Models

Page 14: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Parametric to Geometric[Osher & Sethian 1988]

Level Set PDE:

Contour Deformation:

Page 15: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Visual Concept of GDM

Page 16: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Properties of GDMs• Advantages:

– Produce closed, non-self-intersecting contours– Independent of contour parameterization– Easy to implement: numerical solution of PDEs on

regular computational grid– Stable computations– Automatically changes topology

• Potential disadvantage:– Does not maintain topology

Page 17: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

• GDM cannot control topology• TGDM (ours) preserves topology

Topology Behavior

GDM: Standard GeometricDeformable Model

TGDM: Topology-preserving Geometric Deformable Model

Page 18: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Why Maintain Topology?

GDM: Standard GeometricDeformable Model

TGDM: Topology-preserving Geometric Deformable Model

Page 19: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 20: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Marching Cubes Isosurface• Where is the boundary

defined by a level set function?

• Consider voxel values on corners of a cube

• Label as– above isovalue– below isovalue

• Determine position of triangular mesh surface passing through the cubes by linear interpolation

> 0.5

< 0.5

Voxel values

Page 21: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Digital Connectivity

• Consistent pairs: (foreground,background) → (6,18), (6,26), (18,6), (26,6)

6-connectivity

18-connectivity 26-connectivity

Page 22: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Digital Embedding of Contour Topology

White Points:

Black Points:

• Contour topology is determined by signs of the level set function at pixel locations

• Topology of the implicit contour is the same as the topology of the digital object

Page 23: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Connectivity Rule of Contour

• Topology of digital contour determined by connectivity rule

Same digital object, different topologies

Page 24: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Topology Preservation Principle

• Preserving surface topology is equivalent to maintaining the topology of the digital object

• The digital object can only change topology when the level set function changes sign at a grid point

• Which sign changes can be allowed, and which cannot?

• To prevent the digital object from changing topology, the level set function should only be allowed to change sign at simple points

[Han et al., PAMI, 2003]

Page 25: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Simple Point• Definition: a point is simple if adding or removing

the point from a binary object will not change the object topology

• Determination: can be characterized locally by the configuration of its neighborhood (8- in 2D, 26- in 3D) [Bertrand & Malandain 1994]

SimpleNon-

Simple

Page 26: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

x is a Simple Point

0)( x

x

0)( x

xx

Page 27: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

x is Not a Simple Point

0)( x 0)( xX

X

Page 28: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Topology Preserving Geometric Deformable Model (TGDM)

• Evolve level set function according to GDM• If level set function is going to change sign, check

whether the point is a simple point– If simple, permit the sign-change– If not simple, prohibit the sign-change (replace the grid value by epsilon with same sign)– (Roughly, this step adds 7% computation time.)

• Extract the final contour using a connectivity consistent isocontour algorithm

Page 29: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Ambiguous Faces

Two possible tilings:

Page 30: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Ambiguous Cubes

Two possible tilings:

Page 31: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Connectivity Consistent MC Algorithm

• (black,white)• (18,6) choose b, f• (26,6) choose b, e

(a) (b) (c)

(d) (e) (f)

AmbiguousFace

AmbiguousCube

• (6,18) choose c, f• (6,26) choose c, f

Page 32: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Nested Deformable Surfaces

Pial Surface

Inner Surface

Central Surface

TGDM-3

Initial WM Isosurface

TGDM-2TGDM-1

Page 33: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

TGDM for Inner Surface

Initial WM Isosurface Evolving GM/WM Interface

[Han et al., NeuroImage, 2004]

Page 34: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

IACL

TGDM for Central Surface

Initialize with GM/WM surface Evolving toward Central Surface

Page 35: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

TGDM for Outer Surface

Evolving toward Outer SurfaceStart from Central Surface

Page 36: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Results—Visual Inspection

Sagittal

• surfaces overlaid on cross-sections of the original image

Axial

Page 37: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 38: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Multiple Object Challenges1. Maintenance of

multiple level sets

2. Maintenance of object’s individual topologies and relationships between objects

Anatomical parcellation is not arbitrary

Page 39: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Prior Strategies• N level set functions for N objects

[Paragios00, Brox06, …]– Pros: Flexibility between objects– Cons: Objects might overlap or form gaps;

large memory and computational demands

• Multi-phase [Vese02] (4-color theorem)– Pros: Log(N) level set functions for N

objects; low computational complexity; no overlaps or gaps;

– Cons: Forces limited to region and length terms; little control over individual object forces; no 3D equivalent

Page 40: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Principle of MGDM• Simple point criterion can

be replaced by digital homeomorphism criterion

• Movement of collection of objects occurs primarily at: – edges between two objects

or – junctions between three

objects

• Higher-order relationships can be ignored

Objects are not digitally homeomorphic.

Fan et al., CVPR’08, MMBIA’08

Page 41: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Level Set Function Decomposition• N objects Oi, i=1,…N• Distance to objects:

• Label functions:

L0 = Object

L1 = Nearest neighbor

L2 = 2nd nearest neighbor

Page 42: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Distance and Level Set Functions• Distance-based functions:

• Reconstruction of level set functions:

0(x)

1(x)

2(x)

^

Approximation: valid assuming 3 objects max per junction

2D

3D

Page 43: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Evolution• Recall GDM:

• Required evolutions:

• Distance-based functions:

Page 44: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

• Assume1. Compute forces2. Find “third” neighbor:

3. Compute:

If then setSet

MGDM Algorithm (2D case)4. Compute:

If then setand• Digital topology and

homeomorphism are readily added

Page 45: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Simulation Experiments• Compare algorithms:

– multiphase (MP)– coupled level sets (CLS)– ours (MGDM)

• Objective function (classic Mumford Shah energy; also Chan-Vese for GDM)

• Evaluate:– convergence, memory usage, computation

time, and misclassification percentage

Page 46: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Visual Comparison

Page 47: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Convergence ComparisonE

Iteration

Page 48: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Quantitative Results

Page 49: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Experiment I: Whole Brain SegmentationStructure memberships from TOADS [Bazin 07] (Topology-preserving, Anatomy-driven segmentation)

ii uf 5.0iu

Membership function:

Force:

Balloon force Smooth force

Sulcal CSF

Cerebral GM

Cerebral WM Cerebellar GM

Cerebellar WM

CaudateThalamus PutamenVentriclesBrainstem

Page 50: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Whole Brain Segmentation: 2D Visualization

(a) Original Image (c) No Topology or Smooth

(b) Result from Toads (d) No Topology but Smooth (f) Group Topology

(e) Single Topology

Topology is preserved with DHC

Page 51: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Whole Brain Segmentation: 3D Visualization

Object topology and relationships between objects can be preserved.

(a) No Topology or Smooth (c) Single Object Topology

(b) No Topology but Smooth (d) Group Objects Topology

Page 52: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Experiment II: Thalamic Nuclei Parcellation

MP-RAGE image

Thalamus Membership

TOADS

FA with PEV color map

co-registered

Homogeneous Orientation

Force for the thalamus boundary.

Force for the thalamus nuclei.

Apply to voxels whose label or the first neighbor is the background.

Apply to voxels whose label or the first neighbor belongs to thalamic nuclei.

Different forces applied to different parts of an object

F

Page 53: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Thalamic Nuclei Parcellation

iii xxxf )()(v)( V

Mean principal orientation from region i

)()(5.0)( xxuxf

The force for thalamus nuclei parcellation is designed based on the assumption that the orientations for each nuclei is homogeneous.

The force for thalamus boundary is a combination of balloon and smooth terms.

Membership function of thalamus at voxel x

Page 54: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Thalamic Nuclei Parcellation: Result

(b) Initialization (d) Result(c) Principal Orientation of Thalamus

(a) Membership Function for Thalamus

Front View Off Diagonal ViewOff Left ViewBack View

Page 55: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Outline• Introduction• Deformable models• TGDM: topology-preserving geometric

deformable model• MGDM: multi-object geometric deformable

model• Conclusion

Page 56: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

General Principles• Object topology can be strictly preserved in

geometry deformable models: TGDM• Multiple objects can be simultaneously

segmented and– topology preserved– object relationships preserved– memory efficient– all conventional forces can be applied– guaranteed to have no overlaps or gaps

Page 57: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Remaining Concerns and the Future• How to establish initial object or collection?

– digital topology is not always preserved under simple transformations such as rotation

– recent work on manual skeleton is promising, but tedious

– automatic topology correction is known only for spherical topology, and it is not globally optimum

• Problem of objects getting “stuck”– not so bad in TGDM– much worse in MGDM

Page 58: Image Segmentation of Multiple Objects and Their Compartments Jerry L. Prince Image Analysis and Communications Laboratory (IACL)

Questions?