2d phase unwrapping
Post on 05-Apr-2018
234 Views
Preview:
TRANSCRIPT
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 1/11
Two-Dimensional
Phase UnwrappingUsing Neural Networks
Wade Schwartzkopf, Thomas E. Milner, Joydeep Ghosh,
Brian L. Evans, and Alan C. Bovik
/DERUDWRU\IRU,PDJHDQG9LGHR(QJLQHHULQJ/DERUDWRU\IRU,PDJHDQG9LGHR(QJLQHHULQJ
'HSDUWPHQWRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ'HSDUWPHQWRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ
7KH8QLYHUVLW\RI7H[DVDW$XVWLQ$XVWLQ7;7KH8QLYHUVLW\RI7H[DVDW$XVWLQ$XVWLQ7;
,(((6RXWKZHVW6\PSRVLXPRQ
,PDJH$QDO\VLVDQG,QWHUSUHWDWLRQ
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 2/11
Phase-Based Imaging Methods
SyntheticAperture Radar
Mountains
MagneticResonance
Imaging
Knee
Optical DopplerTomography
Blood Vessel
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 3/11
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 4/11
y(n)+2π
y(n)
y(n)-2π
y(n-1)+π
y(n-1)-π
1-D Phase Unwrapping
• Use information from neighboring pixels
• Errors propagate
Original Signal
0
π
-π
Wrapped Signal
0
π
-π
Unwrapped Signal
0
π
3π
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 5/11
0.0 0.6π
1.6π 1.2π
2-D Phase Unwrapping
• Possible to use second dimension to recover
• Residues: conflicts in different dimensions
0.0 0.6π
-0.4π 1.2π
0.0 0.6π1.6π 1.2πTwo of the
possible orderings
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 6/11
Why Neural Networks?
• Fast unwrapping of ODT velocity images– 10 100 × 100 images per second
– Global optimization methods impractical
– Fast local methods necessary
• Learn features in ODT velocity images of skintissue– Sparse images
– All blood flow follows parabolic law
• Goal: Train neural networks to detect phase
jumps in ODT phase images
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 7/11
Neighboring pixels
Current pixel
Phase Jump Detector
• Feedforward multilayer
perceptron network
• Wrapped values of 5×5
neighborhood of pixels
as inputs to network
• Three outputs for the current
pixel– Positive jump of 2π
– No jump of 2π
– Negative jump of 2π
• Non-iterative, pixel parallel
computation
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 8/11
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 9/11
Simulated Image
Simulated flow in a blood
vessel without wrapping
Simulated flow in a blood
vessel after wrapping
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 10/11
Phase Jump Detection
Location of positive (black) and
negative (white) phase jumps
Pixels in the range [π, 3π)
• Unwrap along columns– Top pixel is always zero since no movement at edge of skin
– At positive (negative) jump, add (subtract) 2π to rest of column
7/31/2019 2d phase unwrapping
http://slidepdf.com/reader/full/2d-phase-unwrapping 11/11
top related