13
Tomographic Image Reconstruction by Total-Variation Minimization Zhifei Zhang Zhifei Zhang @ EECE, UTK 1

PowerPoint Presentation...Title PowerPoint Presentation Author Zhifei Zhang Created Date 4/23/2015 12:50:30 AM

  • Upload
    others

  • View
    2

  • Download
    0

Embed Size (px)

Citation preview

  • Tomographic Image Reconstruction by Total-Variation Minimization

    Zhifei Zhang

    Zhifei Zhang @ EECE, UTK 1

  • Zhifei Zhang @ EECE, UTK 2

    Outline

    1. Learned in class

    2. Total variation

    3. Comparison

  • Zhifei Zhang @ EECE, UTK 3

    Filtered Back Projection (FBP)

  • Zhifei Zhang @ EECE, UTK 4

    Least Square (LS)

    Ax = y

    Find x knowing A and y

    minš’™>šŸŽ

    šŸ

    šŸš‘Øš’™ āˆ’ š’š šŸ

  • Zhifei Zhang @ EECE, UTK 5

    Drawbacks of FBP and LS

    LS:FBP:ā€¢ Need enough projections

    ā€¢ Sharp edge or noise

    ā€¢ Fit noise (over-fitting)

    ā€¢ Sensitive to outliers

    No noise No noiseOriginal SNR=20dB

    ReconstructReconstruct

    FBP LS

  • Zhifei Zhang @ EECE, UTK 6

    What can Total Variation Achieve?

    Original

    No noise SNR=20dB SNR=10dB

    LS

    TV

    Reconstruct

  • Zhifei Zhang @ EECE, UTK 7

    What is Total Variation (TV)?

    124 100 ā‹Æ69 ā‹Æ ā‹Æā‹® ā‹Æ ā‹Æ š‘›Ć—š‘›

    124 100 ā‹Æ69 ā‹Æ ā‹Æā‹® ā‹Æ ā‹Æ š‘›Ć—š‘›

    š‘»š‘½ =

    š’Š,š’‹=šŸ

    š’

    š‘°š’Šš’‹ āˆ’ š‘°š’Š,š’‹+šŸ + š‘°š’Šš’‹ āˆ’ š‘°š’Š+šŸ,š’‹

    TV

  • Zhifei Zhang @ EECE, UTK 8

    What is Total Variation (TV)?

    If let TV approach zero

    Original TV = 0TV = high TV = low

  • Zhifei Zhang @ EECE, UTK 9

    Total-Variation Minimization

    argminš’™>šŸŽ

    šŸ

    šŸš‘Øš’™ āˆ’ š’š šŸ + š€ āˆ™ š‘»š‘½ (š›Œ > šŸŽ)

    ProjectionSystem model Penalty parameter

    WANTED

    š‘»š‘½ = š’Š=šŸ

    š’Ž

    š’‹=šŸ

    š’

    š›š’™š’Šš’‹ šŸš›š’™š’Šš’‹ denotes gradient

    approximation for pixel ij

  • Zhifei Zhang @ EECE, UTK 10

    Comparison (Less Projections)

    TV

    LS

    200 50 20 10 5

  • Zhifei Zhang @ EECE, UTK 11

    TV

    FBP

    Comparison (Less Projections)

    64 128 256 512

  • Zhifei Zhang @ EECE, UTK 12

    Conclusion

    ā€¢ Image reconstruction by TV minimization

    ā€¢ Make image more smooth and edge clear

    ā€¢ Robust to noise and need less projections

    ā€¢ It may be tricky to set the TV parameter š€(get balance between removing noise and preserving details)

  • Zhifei Zhang @ EECE, UTK 13

    Thank You!

    ā€¢ Hansen, Per Christian, and Jakob Heide JĆørgensen. "Total Variation and Tomographic Imaging from Projections." 36th Conference of the Dutch-Flemish Numerical Analysis Communities.

    ā€¢ Dahl, Joachim, et al. "Algorithms and software for total variation image

    reconstruction via first-order methods." Numerical Algorithms 53.1 (2010): 67-92.

    References