Abstract

We present a simple and fast algorithm for view synthesis based on the acquisition of four high-resolution oblique images with a conventional widefield microscope. The images are acquired simultaneously using a partitioned aperture add-on. The technique provides physically valid views of thin samples that are transmitting or fluorescent, as demonstrated with biopsied tissue or green fluorescent protein-labeled brain slices. The goal of this technique is to facilitate image interpretation by conferring impressions of depth that are otherwise absent in standard microscope images.

© 2014 Optical Society of America

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).
  2. S. Ullman and R. Basri, IEEE Trans. Pattern Anal. Mach. Intell. 13, 992 (1991).
    [CrossRef]
  3. S. M. Seitz and C. R. Dyer, in Proceedings of IEEE Workshop on Representation of Visual Scenes (in Conjunction with ICCV ’95) (1995), p. 18.
    [CrossRef]
  4. A. Parthasarathy, K. Chu, T. Ford, and J. Mertz, Opt. Lett. 37, 4062 (2012).
    [CrossRef]
  5. R. Barankov and J. Mertz, Opt. Lett. 38, 3961 (2013).
    [CrossRef]
  6. M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
    [CrossRef]
  7. A. Orth and K. Crozier, Opt. Lett. 38, 2666 (2013).
    [CrossRef]
  8. Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
    [CrossRef]
  9. C. Sheppard, J. Opt. Soc. Am. A 21, 828 (2004).
  10. S. Kang and R. Szeliski, Int. J. Comput. Vis. 58, 139 (2004).
    [CrossRef]

2013 (2)

2012 (1)

2009 (2)

M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
[CrossRef]

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

2004 (2)

S. Kang and R. Szeliski, Int. J. Comput. Vis. 58, 139 (2004).
[CrossRef]

C. Sheppard, J. Opt. Soc. Am. A 21, 828 (2004).

2001 (1)

W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).

1991 (1)

S. Ullman and R. Basri, IEEE Trans. Pattern Anal. Mach. Intell. 13, 992 (1991).
[CrossRef]

Barankov, R.

Basri, R.

S. Ullman and R. Basri, IEEE Trans. Pattern Anal. Mach. Intell. 13, 992 (1991).
[CrossRef]

Chu, K.

Crozier, K.

Dyer, C. R.

S. M. Seitz and C. R. Dyer, in Proceedings of IEEE Workshop on Representation of Visual Scenes (in Conjunction with ICCV ’95) (1995), p. 18.
[CrossRef]

Ford, T.

Goldman, Y.

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Kang, S.

S. Kang and R. Szeliski, Int. J. Comput. Vis. 58, 139 (2004).
[CrossRef]

Levoy, M.

M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
[CrossRef]

McDowall, I.

M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
[CrossRef]

McKenna, J.

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Mertz, J.

Murry, J.

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Orth, A.

Ostap, E.

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Parthasarathy, A.

Schaefer, L.

W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).

Seitz, S. M.

S. M. Seitz and C. R. Dyer, in Proceedings of IEEE Workshop on Representation of Visual Scenes (in Conjunction with ICCV ’95) (1995), p. 18.
[CrossRef]

Sheppard, C.

Sun, Y.

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Swedlow, J.

W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).

Szeliski, R.

S. Kang and R. Szeliski, Int. J. Comput. Vis. 58, 139 (2004).
[CrossRef]

Ullman, S.

S. Ullman and R. Basri, IEEE Trans. Pattern Anal. Mach. Intell. 13, 992 (1991).
[CrossRef]

Wallace, W.

W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).

Zhang, Z.

M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
[CrossRef]

BioTechniques (1)

W. Wallace, L. Schaefer, and J. Swedlow, BioTechniques 31, 1076 (2001).

IEEE Trans. Pattern Anal. Mach. Intell. (1)

S. Ullman and R. Basri, IEEE Trans. Pattern Anal. Mach. Intell. 13, 992 (1991).
[CrossRef]

Int. J. Comput. Vis. (1)

S. Kang and R. Szeliski, Int. J. Comput. Vis. 58, 139 (2004).
[CrossRef]

J. Microsc. (1)

M. Levoy, Z. Zhang, and I. McDowall, J. Microsc. 235, 144 (2009).
[CrossRef]

J. Opt. Soc. Am. A (1)

Nano Lett. (1)

Y. Sun, J. McKenna, J. Murry, E. Ostap, and Y. Goldman, Nano Lett. 9, 2676 (2009).
[CrossRef]

Opt. Lett. (3)

Other (1)

S. M. Seitz and C. R. Dyer, in Proceedings of IEEE Workshop on Representation of Visual Scenes (in Conjunction with ICCV ’95) (1995), p. 18.
[CrossRef]

Supplementary Material (3)

» Media 1: AVI (4072 KB)     
» Media 2: AVI (1259 KB)     
» Media 3: AVI (1469 KB)     

Cited By

OSA participates in CrossRef's Cited-By Linking service. Citing articles from OSA journals and other participating publishers are listed here.

Alert me when this article is cited.


Figures (6)

Fig. 1.
Fig. 1.

Obj, objective lens; TL, tube lens; EL, entrance lens of the partitioned aperture sub-system (focal length fe); PL, PAW lens (focal length fp); Cam, camera. The PAW lens provides simultaneous imaging from four different view angles.

Fig. 2.
Fig. 2.

Schematic of partitioned aperture. Solid line, PAW lens array; dashed lines, projection of the objective aperture onto the lens array. For simplicity we assume a square aperture. The effective NA for each quadrant image is NAp.

Fig. 3.
Fig. 3.

Transmission image of a H&E stained histology slide of healthy colon captured with the partitioned aperture microscope. The PAW lens projects four oblique detection images. These are cropped to 768×768 and co-registered using a sub-pixel correlation procedure (see text). Insets: expanded regions of interest. Parallax is manifested as displacements of out of focus objects along diagonals. Displacements as large as 5 pixels were observed in this example. Scale bar 30 μm.

Fig. 4.
Fig. 4.

PSF density plots. (a) PSF0. (b) PSFT. Axes in units of xλ=λ/2NAp and zλ=λ/NAp2 (for ease of visibility, horizontal lines have been individually autoscaled). (c) Comparison of the PSF widths. Plots of W0 (solid blue) and WT (dashed red). Inset: plot of the ratio WT/W0.

Fig. 5.
Fig. 5.

Brightfield images extracted from a sequence of synthesized views (animation in Media 1). The sample is a H&E stained histology slide of Stage 1, Colon Adenocarcinoma. From left to right, views for θx between 15° and 15°, respectively. None of the views coincide with the raw quadrants. Enlarged areas show that image quality is largely preserved in the synthesized views. Arrows point to locations where the parallax effect is visible. Scale bar 20 μm.

Fig. 6.
Fig. 6.

Single-view excerpts from sequences of synthesized views. (a) Widefield fluorescence imaging of mouse kidney (animation in Media 2). (b) Widefield fluorescence imaging of microvasculature in a mouse brain (animation in Media 3). Scale bars 20 μm.

Equations (18)

Equations on this page are rendered with MathJax. Learn more.

PSF1(x,y,z)=PSF0(xθpz,yθpz,z),
PSF2(x,y,z)=PSF0(xθpz,y+θpz,z),
PSF3(x,y,z)=PSF0(x+θpz,y+θpz,z),
PSF4(x,y,z)=PSF0(x+θpz,yθpz,z),
PSFθ(x,y,z)=PSF0(xθxz,yθyz,z).
PSFT=(PSF1+PSF2+PSF3+PSF4)/4,
PSFx=(PSF1+PSF2PSF3PSF4)/4,
PSFy=(PSF1PSF2PSF3+PSF4)/4.
PSFT=PSF0,
PSFx=θpzPSF0x,
PSFy=θpzPSF0y.
PSFθPSFT+αxPSFx+αyPSFy,
IT=(I1+I2+I3+I4)/4,
Ix=(I1+I2I3I4)/4,
Iy=(I1I2I3+I4)/4,
IθIT+αxIx+αyIy.
Iθ=[(1+αx)(1+αy)I1+(1+αx)(1αy)I2+(1αx)(1αy)I3+(1αx)(1+αy)I4]/4.
W(z)=(PSFdxdy)2/PSF2dxdy.

Metrics