OSA's Digital Library

Biomedical Optics Express

Biomedical Optics Express

  • Editor: Joseph A. Izatt
  • Vol. 5, Iss. 3 — Mar. 1, 2014
  • pp: 907–920
« Show journal navigation

Isotropic scalar image visualization of vector differential image data using the inverse Riesz transform

Kieran G. Larkin and Peter A. Fletcher  »View Author Affiliations


Biomedical Optics Express, Vol. 5, Issue 3, pp. 907-920 (2014)
http://dx.doi.org/10.1364/BOE.5.000907


View Full Text Article

Acrobat PDF (5544 KB)





Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

X-ray Talbot moiré interferometers can now simultaneously generate two differential phase images of a specimen. The conventional approach to integrating differential phase is unstable and often leads to images with loss of visible detail. We propose a new reconstruction method based on the inverse Riesz transform. The Riesz approach is stable and the final image retains visibility of high resolution detail without directional bias. The outline Riesz theory is developed and an experimentally acquired X-ray differential phase data set is presented for qualitative visual appraisal. The inverse Riesz phase image is compared with two alternatives: the integrated (quantitative) phase and the modulus of the gradient of the phase. The inverse Riesz transform has the computational advantages of a unitary linear operator, and is implemented directly as a complex multiplication in the Fourier domain also known as the spiral phase transform.

© 2014 Optical Society of America

1. Introduction

Recent advances in X-ray imaging allow the capture of differential phase images of the test specimen. Differential phase provides information about X-ray deflection rather that the X-ray absorption of conventional radiography. Clinical differential phase provides fine detail about the soft-tissue structure, detail simply not present in normal X-ray images. However, there is a catch: differential phase gives a highly asymmetric and anisotropic depiction of the optical path (or integrated refractive index) through a specimen. Many systems (using linear gratings) provide just one oriented differential phase image - the x derivative for example. More recent systems have been developed with crossed-gratings (or 2-D arrays) that simultaneously provide two differential phase images; the x and the y derivatives for example.

  • integration may encounter inconsistent gradient data between two components, resulting in damaging image artifacts,
  • integration typically increases dynamic range beyond the range of LCD displays,
  • integration introduces blur due to the intrinsic low-pass nature of 2-D integrator.

In this paper we propose extending the 1-D Hilbert symmetrizing method formally for 2-D images, using the Riesz transformation. Only the essential underlying mathematical theory will be presented in this preliminary exposition. The paper is structured as follows:

  • Section 2 outlines the Fourier theory of differential imaging.
  • Section 3 introduces the Riesz transform and the related complex embedding (known as the spiral phase transform)
  • Section 4 outlines the main algorithmic definition of the process.
  • Section 5 presents inverse Riesz visualization in comparison with conventional images.
  • Section 6 concludes with a discussion of the main results.

2. Differential phase imaging prerequisites

For convenience we consider the differential phase images generated by a 2-D X-ray grating-based phase contrast imaging system [3

3. H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express 19(4), 3339–3346 (2011). [CrossRef] [PubMed]

]. The essential principal behind the 1-D Talbot system was first proposed by Momose et al [4

4. A. Momose, S. Kawamoto, I. Koyama, Y. Hamaishi, K. Takai, and Y. Suzuki, “Demonstration of X-ray Talbot interferometry,” Jpn. J. Appl. Phys., Part 2 , 42, 866–868 (2003).

], then refined by Pfeiffer et al [5

5. F. Pfeiffer, T. Weitkamp, O. Bunk, and C. David, “Phase retrieval and differential phase-contrast imaging with low-brilliance X-ray sources,” Nat. Phys. 2(4), 258–261 (2006). [CrossRef]

]. The 2-D system multiplexes the different directional phase derivatives onto different Fourier spectral side-lobes. Alternatively, we could just assume that the two requisite differential phase images are provided from a black-box system (such as two separate, orthogonal, Nomarski DIC images).

2.1 Differential phase from 2-D X-ray grating Fourier side-lobes

First we define the spatial and Fourier coordinate systems, both Cartesian and polar:
(x,y),(r,θ),x=rcosθ,y=rsinθ(u,v),(q,ϕ),u=qcosϕ,v=qsinϕ}
The simplest system generates a moiré intensity image f(x,y) on the sensor with nine Fourier spectral lobes, as follows:
f(x,y)=n=1+1m=1+1bmn(x,y)exp[2πi(mu0+nv0)+i[mψx(x,y)+nψy(x,y)]]
(1)
The intensity image is real and positive. Figure 1(a)
Fig. 1 (a) Simulation of a deeply phase modulation moiré pattern f(x,y), and (b) its Fourier transform magnitude |F(u,v)| clearly showing the DC lobe and eight side-lobes.
shows a simulated moiré pattern for an exaggerated, but simple object. The nine separate lobes are clear.

We denote the 2-D continuous Fourier transform (FT) as a double headed arrow F(u,v)FTf(x,y)and define it via the forward and inverse transforms:
F(u,v)=++f(x,y)exp[2πi(ux+vy)]dxdyf(x,y)=++F(u,v)exp[+2πi(ux+vy)]dudv}
(2)
In practice we utilize discrete Fourier transforms (DFTs) and in particular fast Fourier transforms (FFTs), but for simplicity descriptions are in terms of continuous transforms. As the intensity in Eq. (2) is real, the FT of the moiré image in Eq. (2) is Hermitian, namely the sum of nine triple convolutions which correspond to four Hermitian pairs and a DC lobe:

F(u,v)=n=1+1m=1+1Bmn(u,v)Pmx(umu0,v)Pny(u,vnv0)
(3)

2.2 Phase gradients and Fourier correspondence

3. Riesz transform

The Riesz transform is widely accepted in the mathematics community (harmonic theory in particular) as the proper, isotropic generalization of the 1-D Hilbert transform to higher dimensions (or Euclidean spaces). In engineering, especially signal and image processing, anisotropic approximations like the orthant Hilbert transform remain popular, and the Riesz transform is little used.

3.1 Origin of the Riesz transform

We have already stated the Fourier multiplier property of the gradient operator in Eq. (9). It is not difficult to prove by differentiation of Eq. (2):
xf(x,y)=++F(u,v)2πiuexp[+2πi(ux+vy)]dudvyf(x,y)=++F(u,v)2πivexp[+2πi(ux+vy)]dudv}
(10)
Similarly the Laplacian operator Δ=div.grad has the following form:
Δ=2x2+2y2Δf(x,y)=++F(u,v)[(2πiu)2+(2πiv)2]exp[+2πi(ux+vy)]du}
(11)
Written in more expressive shorthand we have:
Δf(x,y)FT(2π)2(u2+v2)F(u,v)=(2πq)2F(u,v)
(12)
It looks like the Laplacian operator has a possible “square root” in the Fourier domain. The square root of the Laplacian occurs in the mathematical foundations of tomography [15

15. R. S. Strichartz, “Radon inversion - variations on a theme,” Am. Math. Mon. 89(6), 377–384 (1982). [CrossRef]

]. For example, Smith & Keinert [16

16. K. T. Smith and F. Keinert, “Mathematical foundations of computed tomography,” Appl. Opt. 24(23), 3950–3957 (1985). [CrossRef] [PubMed]

] define it as their Λ (lambda) operator:
(Δ)12f(x,y)=Λf(x,y)FT2πqF(u,v)
(13)
Faridani & Smith [17

17. A. Faridani, E. L. Ritman, and K. T. Smith, “Local tomography,” SIAM J. Appl. Math. 52(2), 459–484 (1992). [CrossRef]

] then use the lambda operator defined above to generate a locally soluble tomographic inversion based on the known spatial localization of the Laplacian operator. In contrast our proposed application maintains the spatial non-locality of the Riesz operators. We factorize the gradient operator, applied to the phase in this instance, as follows:
ψ=ΛRψFT2πiu2+v2[ui+vj]u2+v2Ψ(u,v)
(14)
The new vector operator R is the Riesz transform, and its Fourier multiplier has unit magnitude and odd symmetry, very like the Hilbert multiplier, and can be compactly expressed in terms of the Fourier polar angle ϕ:
Rψ=(Δ)12ψFTi[ui+vj]u2+v2Ψ(u,v)=i(icosϕ+jsinϕ)Ψ(u,v)
(15)
The Riesz transform has numerous useful and special properties that make it invaluable in proving important theorems in the foundations of harmonic analysis, but brevity prohibits further elaboration.

3.2 Embedding of the vector Riesz transform in the complex scalar domain

It often happens that the Riesz transform needs to be implemented in a pre-existing image processing software application. Many applications work with gray-scale images and allow complex representation for compatibility with discrete Fourier transform (DFT) processing. Surprisingly, Riesz transform processing can be implemented in 2-D without the further extension to complex vector representation. A little more care has to be taken to account for cross-talk between somewhat overloaded channels, but in most cases everything just works out automatically. The embedded formalism is known as the complexified Riesz transform or spiral phase transform [13

13. K. G. Larkin, D. J. Bone, and M. A. Oldfield, “Natural demodulation of two-dimensional fringe patterns. I. General background of the spiral phase quadrature transform,” J. Opt. Soc. Am. A 18(8), 1862–1870 (2001). [CrossRef] [PubMed]

].

3.3 Discrepancy term of the inverse Riesz transform

Considering Eq. (19) it would seem that application of the properly scaled inverse Riesz transform results in a real image with lambda enhancement. However we have not considered random noise or systematic effects. In general each of the differential images will be a pure differential image with some unknown additive component. From our work [20

20. K. G. Larkin and P. A. Fletcher, “A coherent framework for fingerprint analysis: are fingerprints Holograms?” Opt. Express 15(14), 8667–8677 (2007). [CrossRef] [PubMed]

] in general decompositions of 2-D phase fields and 2-D phase unwrapping we recognized immediately that there is a part of the phase that is not representable as the gradient of a scalar. This so-called phase “discrepancy” can be represented as the curl of a vector field, and leads inevitably from the Helmholtz decomposition (viz the Fundamental Theorem of Vector Calculus). Ghiglia and Pritt (chapter 2 [6

6. D. C. Ghiglia and M. D. Pritt, Two-dimensional phase unwrapping (John Wiley and Sons, New York, 1998).

],) describe in detail the conditions necessary for a phase function in 2-D to generate a curl discrepancy. Even for non-phase differentials, noise can still induce apparent discrepancies.

One important systematic error can easily occur if the direction of the derivative operation is rotated. For example, if the x differential is really along a direction rotated an angle β counterclockwise from the x-axis, and the y differential an angle β counterclockwise from the y-axis, then the output of the inverse Riesz defined in Eq. (19) is simply multiplied by a complex constant exp[iβ]. The end result is that if the x and y differentials are interchanged, for example, then the desired output switches from the real part to the imaginary part. So it is important to make sure that the derivative direction is known beforehand. In practice it is possible to automate algorithms to compensate this effect under normal circumstances (when the noise level is not excessive).

Note that the rotation of derivative directions gives the same effect as introducing a curl (discrepancy) component, and this fact ties in with the observation that the curl effect can be interpreted as 90º rotated gradient components.

More generally, the discrepancy term is a measure of how reliable the gradient data are. The conventional approach to integration by solving the Poisson equation is completely blind to the discrepancy data. Discrepancy is discussed further in section 5.4.

4. Algorithmic definition inverse Riesz transform applied to differential images

As outlined above, the complexified Riesz transform, and inverse Riesz transform are simple to implement in an image processing system that allows complex variables. Figure 2
Fig. 2 Flowchart for inverse Riesz transformation of gradient components
shows the basic flow chart.

Two orthogonal differential images are input into the processor. The x-differential image is placed in the real part, and the y-differential in the imaginary part. The fast Fourier transform of the complexified gradient image is computed and the result is then multiplied by the inverse Riesz spiral phase factor. Complex constants 2πi are used to re-normalize the gradient Fourier multiplier. However this step may be omitted as it only multiplies the final image by a constant factor and swaps the real and imaginary parts. The inverse FFT (IFFT) produces an output with the desired image in the real part and a “discrepancy” image in the imaginary channel.

5. Application to differential images

Although our proposed method is applicable to any differential image pairs, whether phase or otherwise, we have chosen an example from X-ray Talbot moiré differential phase imaging.

5.1 Example from X-ray Talbot moiré interferometry

Our example is a molded plastic chess piece shown in Fig. 3
Fig. 3 The test specimen is a plastic chess piece (a knight).
. The chess-piece was imaged in an experimental X-Ray Talbot moiré interferometer [3

3. H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express 19(4), 3339–3346 (2011). [CrossRef] [PubMed]

]. Some plastics have low X-ray absorption yet significant refractive index not unlike some animal soft tissues. A complete sequence of 16 (4x4) phase shifted interferograms was collected. Each phase shift is 90º in a predominantly horizontal or vertical direction. One frame of the sequence is shown in Fig. 4
Fig. 4 One of sixteen phase-shifted moiré interferograms, f(x,y), from an experimental X-ray Talbot grating interferometer . The specimen is a plastic chess piece (a knight). The image is 1024 x 1024 pixels, pixel size 48μm.
.

Two Fourier side-lobes were isolated using 16 step 2-D phase shifting algorithms (PSAs). The side-lobes could easily have been isolated using the windowed Fourier transform method [21

21. K. Nagai, H. Itoh, G. Sato, T. Nakamura, K. Yamaguchi, T. Kondoh, and S. H. T. Den, “New phase retrieval method for single-shot x-ray Talbot imaging using windowed Fourier transform,” Optical Modeling and Performance Predictions V, Proc. SPIE 8127, San Diego, California, August 21, 2011.

], or the traditional Fourier transform method [22

22. D. J. Bone, H.-A. Bachor, and R. J. Sandeman, “Fringe-pattern analysis using a 2-D Fourier transform,” Appl. Opt. 25(10), 1653–1660 (1986). [CrossRef] [PubMed]

,23

23. M. Takeda, H. Ina, and S. Kobayashi, “Fourier-transform method of fringe-pattern analysis for computer-based topography and interferometry,” J. Opt. Soc. Am. 72(1), 156–160 (1982). [CrossRef]

], but we prefer to illustrate our method with the highest isolation, resolution and noise reduction obtained from a PSA.

Figure 5(a)
Fig. 5 (a) Fourier transform magnitude |F(u,v)|, (log-scale). (b) Absorption image b00 (x,y) (Media 1)
shows the corresponding Fourier transform magnitude, the extracted amplitude modulation (or absorption) image, and the extracted x and y differential phase images. The absorption image corresponds to the conventional X-ray image. The variation in background intensity of the absorption in Fig. 5(b) is due to the interferometer source geometry [3

3. H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express 19(4), 3339–3346 (2011). [CrossRef] [PubMed]

], but has no effect on the phase estimation.

The x and y differentials in Fig. 6(a)
Fig. 6 (a) X-differential phase image ψx(x,y) (Media 2), (b) Y-differential phase image ψy(x,y) (Media 3)
and 6(b) have the well-known bas-relief asymmetry of optical differential phase images. On their own the differential images are highly anisotropic, losing considerable detail at certain orientations. Together the images are difficult to interpret; requiring continual visual scanning and comparison.

5.2 Visual assessment

In Fig. 7(a) we see that integration produces low frequency background variation of intensity (mainly due to our imposing overly simplistic periodic boundary conditions). Furthermore the integrated image has a large dynamic range (hard to view properly on a monitor or in print). There is also a low-pass filtering effect apparent when compared to the original differentials.

The gradient magnitude images of Fig. 7(b) and 7(c) seem to maintain the required detail and isotropy with the negated magnitude of Fig. 7(b) suitable for print out and Fig. 7(c) perhaps more suitable for monitor display. What is not so obvious in this example is the potential aliasing that the modulus operation generates. It is well known in signal processing that modulus-squaring doubles the bandwidth required by an image without generating any more useful information (quite the contrary in fact). Furthermore the modulus operation generates even higher order aliasing than mod-squared. The absorption image and the integrated image both have the advantage that the (demodulation) processing is linear (in terms of the 16 input moiré images); and so no spurious (aliased) frequencies are generated.

The inverse Riesz transformed image in Fig. 7(d) we see an image with a wealth of fine detail. The image also retains some of the global intensity ordering of the idealized integral image in Fig. 7(a). Dark regions in the inverse Riesz image broadly correspond with dark regions in the integrated image – so intensity ordering is approximately maintained. Although the gradient image of Fig. 7(b) has fine detail it has completely lost the intensity ordering. For example all the air bubbles are rendered as brightly as their surrounding regions. The inverse Riesz image of Fig. 7(d) shows the classic edge enhancement common to optical phase contrast imaging systems. In this instance the edge effect results from two processes, the first being the differential phases from Talbot distance propagation in the interferometer, the second from the isotropic melding of the inverse Riesz transformer. However, the overshoot/undershoot is purely a mathematical characteristic of the lambda operator. Although we do not currently have a quantitative criterion for comparing the relevant information content of Fig. 7 (a), Fig. 7(b) and Fig. 7(d) we think that our initial visual comparison will be compelling for many readers. The main properties can be summarized in a Table 1

Table 1. Comparative properties of imaging modes*

table-icon
View This Table
:

5.3 Optical (Craik-O'Brien-Cornsweet) illusion

Careful measurement of local patch intensities shows that the intensity ordering in the Riesz image is to some extent an optical illusion, albeit serendipitous. For example the internal air bubble near the middle of the bottom edge appears dark in both Fig. 7(a) and 7(d). One of the reasons the Riesz processed image resembles the integrated image is that the sharp overshooting edges produce apparently darker or lighter regions via a human perceptual effect, known as the Craik-O'Brien-Cornsweet illusion [24

24. M. W. Levine and J. M. Shefner, Fundamentals of Sensation and Perception (Pacific Grove, CA: Brooks/Cole, 1991) p. 675.

]. The quantitatively darker regions in the integral image coincide with the apparently darker regions in the Riesz image, likewise for the lighter regions. Adjacent regions with the same intensity will be perceived to be different if the connecting edges have an overshooting and undershooting intensity profile.

5.4 Discrepancy image

6. Discussion

The critical reader might suggest that our proposed method is merely equivalent to proceeding with conventional integration, then applying a high-pass (i.e. lambda or ramp) filter. Whilst this is almost true, our method completely avoids the most difficult and unstable step of integration. The inverse Riesz transform is a stable, unitary linear operation. Conventional (Poisson equation) integration also omits the informative discrepancy image. The Riesz transform can be viewed, then, as a simple short-cut to a rather desirable result.

Acknowledgments

We would like to thank our Canon Inc. colleagues for providing access to an X-ray Talbot moiré interferometer [3

3. H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express 19(4), 3339–3346 (2011). [CrossRef] [PubMed]

], and experimental data summarized by Fig. 4. In particular, we thank Kentaro Nagai, Soichiro Handa, Takashi Nakamura, Takeshi Kondoh, Yutaka Setomoto, Takayuki Teshima, and Hidenosuke Itoh. Don Bone provided the chess piece appearing in Figs. 3-8 and collected the corresponding interferogram data.

References and links

1.

M. R. Arnison, K. G. Larkin, C. J. R. Sheppard, N. I. Smith, and C. J. Cogswell, “Linear phase imaging using differential interference contrast microscopy,” J. Microsc. 214(1), 7–12 (2004). [CrossRef] [PubMed]

2.

M. R. Arnison, C. J. Cogswell, N. I. Smith, P. W. Fekete, and K. G. Larkin, “Using the Hilbert transform for 3D visualization of differential interference contrast microscope images,” J. Microsc. 199(1), 79–84 (2000). [CrossRef] [PubMed]

3.

H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express 19(4), 3339–3346 (2011). [CrossRef] [PubMed]

4.

A. Momose, S. Kawamoto, I. Koyama, Y. Hamaishi, K. Takai, and Y. Suzuki, “Demonstration of X-ray Talbot interferometry,” Jpn. J. Appl. Phys., Part 2 , 42, 866–868 (2003).

5.

F. Pfeiffer, T. Weitkamp, O. Bunk, and C. David, “Phase retrieval and differential phase-contrast imaging with low-brilliance X-ray sources,” Nat. Phys. 2(4), 258–261 (2006). [CrossRef]

6.

D. C. Ghiglia and M. D. Pritt, Two-dimensional phase unwrapping (John Wiley and Sons, New York, 1998).

7.

R. N. Bracewell, The Fourier Transform and its Applications, McGraw-Hill Electrical and Electronic Engineering Series (McGraw Hill, New York, 1978).

8.

K. R. Freischlad and C. L. Koliopoulos, “Modal estimation of a wave front from difference measurements using the discrete Fourier transform,” J. Opt. Soc. Am. A 3(11), 1852–1861 (1986). [CrossRef]

9.

R. T. Frankot and R. Chellappa, “A method for enforcing integrability in shape from shading algorithms,” IEEE Trans. Pattern Anal. Mach. Intell. 10(4), 439–451 (1988). [CrossRef]

10.

D. C. Ghiglia and L. A. Romero, “Robust Two-Dimensional Weighted and Unweighted Phase Unwrapping That Uses Fast Transforms and Iterative Methods,” J. Opt. Soc. Am. A 11(1), 107–117 (1994). [CrossRef]

11.

S. Velghe, J. Primot, N. Guérineau, M. Cohen, and B. Wattellier, “Wave-front reconstruction from multidirectional phase derivatives generated by multilateral shearing interferometers,” Opt. Lett. 30(3), 245–247 (2005). [CrossRef] [PubMed]

12.

E. M. Stein, Singular Integrals and Differentiability Properties of Functions, Princeton mathematical series (Princeton University Press, Princeton, N.J., 1970).

13.

K. G. Larkin, D. J. Bone, and M. A. Oldfield, “Natural demodulation of two-dimensional fringe patterns. I. General background of the spiral phase quadrature transform,” J. Opt. Soc. Am. A 18(8), 1862–1870 (2001). [CrossRef] [PubMed]

14.

M. Felsberg and G. Sommer, “The Monogenic Signal,” IEEE Trans. Signal Process. 49(12), 3136–3144 (2001). [CrossRef]

15.

R. S. Strichartz, “Radon inversion - variations on a theme,” Am. Math. Mon. 89(6), 377–384 (1982). [CrossRef]

16.

K. T. Smith and F. Keinert, “Mathematical foundations of computed tomography,” Appl. Opt. 24(23), 3950–3957 (1985). [CrossRef] [PubMed]

17.

A. Faridani, E. L. Ritman, and K. T. Smith, “Local tomography,” SIAM J. Appl. Math. 52(2), 459–484 (1992). [CrossRef]

18.

M. Reisert and H. Burkhardt, “Complex derivative filters,” IEEE Trans. Image Process. 17(12), 2265–2274 (2008). [CrossRef] [PubMed]

19.

M. Unser and N. Chenouard, “A Unifying Parametric Framework for 2D Steerable Wavelet Transforms,” SIAM J. Imaging Sciences 6(1), 102–135 (2013). [CrossRef]

20.

K. G. Larkin and P. A. Fletcher, “A coherent framework for fingerprint analysis: are fingerprints Holograms?” Opt. Express 15(14), 8667–8677 (2007). [CrossRef] [PubMed]

21.

K. Nagai, H. Itoh, G. Sato, T. Nakamura, K. Yamaguchi, T. Kondoh, and S. H. T. Den, “New phase retrieval method for single-shot x-ray Talbot imaging using windowed Fourier transform,” Optical Modeling and Performance Predictions V, Proc. SPIE 8127, San Diego, California, August 21, 2011.

22.

D. J. Bone, H.-A. Bachor, and R. J. Sandeman, “Fringe-pattern analysis using a 2-D Fourier transform,” Appl. Opt. 25(10), 1653–1660 (1986). [CrossRef] [PubMed]

23.

M. Takeda, H. Ina, and S. Kobayashi, “Fourier-transform method of fringe-pattern analysis for computer-based topography and interferometry,” J. Opt. Soc. Am. 72(1), 156–160 (1982). [CrossRef]

24.

M. W. Levine and J. M. Shefner, Fundamentals of Sensation and Perception (Pacific Grove, CA: Brooks/Cole, 1991) p. 675.

OCIS Codes
(100.3010) Image processing : Image reconstruction techniques
(110.6760) Imaging systems : Talbot and self-imaging effects
(180.3170) Microscopy : Interference microscopy
(340.7450) X-ray optics : X-ray interferometry
(330.4595) Vision, color, and visual optics : Optical effects on vision
(100.4994) Image processing : Pattern recognition, image transforms

ToC Category:
X-Ray Microscopy and Imaging

History
Original Manuscript: December 13, 2013
Revised Manuscript: February 4, 2014
Manuscript Accepted: February 17, 2014
Published: February 26, 2014

Citation
Kieran G. Larkin and Peter A. Fletcher, "Isotropic scalar image visualization of vector differential image data using the inverse Riesz transform," Biomed. Opt. Express 5, 907-920 (2014)
http://www.opticsinfobase.org/boe/abstract.cfm?URI=boe-5-3-907


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. M. R. Arnison, K. G. Larkin, C. J. R. Sheppard, N. I. Smith, and C. J. Cogswell, “Linear phase imaging using differential interference contrast microscopy,” J. Microsc.214(1), 7–12 (2004). [CrossRef] [PubMed]
  2. M. R. Arnison, C. J. Cogswell, N. I. Smith, P. W. Fekete, and K. G. Larkin, “Using the Hilbert transform for 3D visualization of differential interference contrast microscope images,” J. Microsc.199(1), 79–84 (2000). [CrossRef] [PubMed]
  3. H. Itoh, K. Nagai, G. Sato, K. Yamaguchi, T. Nakamura, T. Kondoh, C. Ouchi, T. Teshima, Y. Setomoto, and T. Den, “Two-dimensional grating-based X-ray phase-contrast imaging using Fourier transform phase retrieval,” Opt. Express19(4), 3339–3346 (2011). [CrossRef] [PubMed]
  4. A. Momose, S. Kawamoto, I. Koyama, Y. Hamaishi, K. Takai, and Y. Suzuki, “Demonstration of X-ray Talbot interferometry,” Jpn. J. Appl. Phys., Part 2, 42, 866–868 (2003).
  5. F. Pfeiffer, T. Weitkamp, O. Bunk, and C. David, “Phase retrieval and differential phase-contrast imaging with low-brilliance X-ray sources,” Nat. Phys.2(4), 258–261 (2006). [CrossRef]
  6. D. C. Ghiglia and M. D. Pritt, Two-dimensional phase unwrapping (John Wiley and Sons, New York, 1998).
  7. R. N. Bracewell, The Fourier Transform and its Applications, McGraw-Hill Electrical and Electronic Engineering Series (McGraw Hill, New York, 1978).
  8. K. R. Freischlad and C. L. Koliopoulos, “Modal estimation of a wave front from difference measurements using the discrete Fourier transform,” J. Opt. Soc. Am. A3(11), 1852–1861 (1986). [CrossRef]
  9. R. T. Frankot and R. Chellappa, “A method for enforcing integrability in shape from shading algorithms,” IEEE Trans. Pattern Anal. Mach. Intell.10(4), 439–451 (1988). [CrossRef]
  10. D. C. Ghiglia and L. A. Romero, “Robust Two-Dimensional Weighted and Unweighted Phase Unwrapping That Uses Fast Transforms and Iterative Methods,” J. Opt. Soc. Am. A11(1), 107–117 (1994). [CrossRef]
  11. S. Velghe, J. Primot, N. Guérineau, M. Cohen, and B. Wattellier, “Wave-front reconstruction from multidirectional phase derivatives generated by multilateral shearing interferometers,” Opt. Lett.30(3), 245–247 (2005). [CrossRef] [PubMed]
  12. E. M. Stein, Singular Integrals and Differentiability Properties of Functions, Princeton mathematical series (Princeton University Press, Princeton, N.J., 1970).
  13. K. G. Larkin, D. J. Bone, and M. A. Oldfield, “Natural demodulation of two-dimensional fringe patterns. I. General background of the spiral phase quadrature transform,” J. Opt. Soc. Am. A18(8), 1862–1870 (2001). [CrossRef] [PubMed]
  14. M. Felsberg and G. Sommer, “The Monogenic Signal,” IEEE Trans. Signal Process.49(12), 3136–3144 (2001). [CrossRef]
  15. R. S. Strichartz, “Radon inversion - variations on a theme,” Am. Math. Mon.89(6), 377–384 (1982). [CrossRef]
  16. K. T. Smith and F. Keinert, “Mathematical foundations of computed tomography,” Appl. Opt.24(23), 3950–3957 (1985). [CrossRef] [PubMed]
  17. A. Faridani, E. L. Ritman, and K. T. Smith, “Local tomography,” SIAM J. Appl. Math.52(2), 459–484 (1992). [CrossRef]
  18. M. Reisert and H. Burkhardt, “Complex derivative filters,” IEEE Trans. Image Process.17(12), 2265–2274 (2008). [CrossRef] [PubMed]
  19. M. Unser and N. Chenouard, “A Unifying Parametric Framework for 2D Steerable Wavelet Transforms,” SIAM J. Imaging Sciences6(1), 102–135 (2013). [CrossRef]
  20. K. G. Larkin and P. A. Fletcher, “A coherent framework for fingerprint analysis: are fingerprints Holograms?” Opt. Express15(14), 8667–8677 (2007). [CrossRef] [PubMed]
  21. K. Nagai, H. Itoh, G. Sato, T. Nakamura, K. Yamaguchi, T. Kondoh, and S. H. T. Den, “New phase retrieval method for single-shot x-ray Talbot imaging using windowed Fourier transform,” Optical Modeling and Performance Predictions V, Proc. SPIE 8127, San Diego, California, August 21, 2011.
  22. D. J. Bone, H.-A. Bachor, and R. J. Sandeman, “Fringe-pattern analysis using a 2-D Fourier transform,” Appl. Opt.25(10), 1653–1660 (1986). [CrossRef] [PubMed]
  23. M. Takeda, H. Ina, and S. Kobayashi, “Fourier-transform method of fringe-pattern analysis for computer-based topography and interferometry,” J. Opt. Soc. Am.72(1), 156–160 (1982). [CrossRef]
  24. M. W. Levine and J. M. Shefner, Fundamentals of Sensation and Perception (Pacific Grove, CA: Brooks/Cole, 1991) p. 675.

Cited By

Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.

Supplementary Material


» Media 1: JPG (736 KB)     
» Media 2: JPG (593 KB)     
» Media 3: JPG (578 KB)     
» Media 4: JPG (191 KB)     
» Media 5: JPG (706 KB)     
» Media 6: JPG (706 KB)     
» Media 7: JPG (588 KB)     

« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited