The video is available at: http://www.youtube.com/watch?v=YS9bAWLins4 All logos are trademark of respective owners. They are used here for illustrative purpose. Images acknowledgement: INRIA, INRA, Institut Pasteur, UHA, Wikimedia commons, and others Images are used here for illustration only and is not for commercial interests. Research results are from my own work. Email: praveen.pankaj@gmail.com

1. Hello!
My name is Praveen Pankajakshan, ...
1

2. ... and this is my
Y
2

3. Biography Cloud
3 Powered by: tagul.com

4. Academics
4

5. UNDERGRADUATE
Electrical Engineering
IIT ROORKEE
INDIA
5

6. GRADUATE SCHOOL
Electrical & Computer Engineering
Texas A&M University
College Station
USA
6

7. PHD SUMMA CUM LAUDE
Computer Science
INRIA Sophia Antipolis
&
University of Nice-Sophia Antipolis
7

8. he sis
CORDI-s Fellowship D T port
Ph Re
J ury
8

9. What I do now ...
Post Doctoral Fellow
9

10. PhD+Postdoc research cloud
10 Powered by: academia.edu

11. Where I live ...
x
11

12. Expertise: Computational Nanoscopy
12

13. For me research is like fine cuisine ...
Objectivity Creativity Tools
13

14. My
Experiences
are drawn from
14

15. My diverse research cloud...
15 Powered by: tagul.com

16. Wh
at
Ib
eli
eve
in
. ..
16
Co
lla
Re bor
sea ati
rch ve

17. Upcoming
book chapter*
2012
+
12 peer reviewed
conference and journal
articles
+
many invited talks
*P. Pankajakshan, et al. Deconvolution and Denoising for Confocal Microscopy.
In F.Cazals and P. Kornprobst, editor, Modeling in Computational Biology and
17 Biomedicine, ch.4, Springer, In Publication, 2012.

18. Biography listed in
Who’s who in the World
2010, 2011, 2012
18

19. ’10
LS
YR er Aw ard
ost
PSF model for fluorescence MACROscopy imaging
ic P
Praveen Pankajakshan1, Alain Dieterlen2, Gilbert Engler3, Zvi Kam4,
ubl
Laure Blanc-Féraud5, Josiane Zerubia5 & Jean-Christophe Olivo-Marin1
1
Quantitative Image Analysis Unit, Pasteur Institute, France,
2
P
Laboratoire MIPS-LAB.EL, Universite de Haute-Alsace, France.
3
IBSV Unit, INRA, France,
4
Molecular Cell Biology, Weizmann Institute of Science, Israel.
5
Ariana joint research group, INRIA/CNRS/UNS, France.
INTRODUCTION OBJECTIVE
To model the point-spread function (PSF) of a MACROscope operating with field
Fluorescent MACROscope is useful for observing large samples (of the order of a few
aberrations due to optical vignetting.
centimeters) and has the following advantages:
• large object fields,
• large working distances, and
• parallax-free imaging.
Imaging field aberrations using point source
Telecentric lens assembly MACROscope lens assembly
Why A PSF MODEL IS IMPORTANT?
Z
Z
XY Observed PSF Unknown
XY Point-spread function (PSF)-> Image of ideal point source (4 m).
Total magnification 6.25x. volume synthetic object
Bio ’10 d
METHODOLOGY
Stokseth’s PSF model Excitation PSF Emission PSF
Lens displacement
y-direction
M iFo Aw ar
ter n
Pupil function for a MICROscope
{
Pos iko
{
Pupil function for a MACROscope
Lens displacement
est
x-direction
Results
B N
SIMULATE PUPIL OPTIC
AL
VIGNETTING AND
from
PSF
CONCLUSIONS
Z Cat’s eye 1. MACROscope PSF varies as a function of the lateral position.
2. Vignetting was observed for small zooms (large FOV).
X X
effect
X
XY 3. This initial PSF model will be enhanced with newer
X X
acquisitions on different systems.
Z
XY
ACKNOWLEDGEMENTS
The first author wish to thank ANR DIAMOND for funding the
Simulated PSF postdoctoral research fellowship. The authors also are grateful to
Z
Dr. Philippe Herbomel from the Institute Pasteur, France and Dr.
Z
Didier Hentsch from IGBMC, France for the images and the
XY XY discussions.
ed bead www-syscom.univ-mlr.fr/ANRDIAMOND/
Experimentally obtain
l
images for different latera Measured bead image
www.bioimageanalysis.org
position in the field
{praveen,jcolivo}@pasteur.fr
alain.dieterlen@uha.fr
19

20. If you like my work and have a
X
20

21. Contact me ...
praveen.pankaj@gmail.com Praveen Pankajakshan
+33 (0) 628358064
15 rue de l’ésperance
www.bioimageanalysis.org/~praveen
75013 Paris
www.linkedin.com/in/praveenpankaj
pasteur.academia.edu/PraveenPankajakshan
praveenpankaj France
21

22. Research Highlight
22

23. ed
e rv
O bs
v ed
v ol
on
Dec
Deconvolution as virtual inverse lens
23

24. Application: Fluorescence Microscopy
Widefield (WFM) Confocal (CLSM)











M. Minsky. Memoir on inventing the confocal scanning
microscope. Scanning, 10:128–138, 1988.
24

25. Volume on Volume on
WFM CLSM
Background Better but ...
Fluorescence! Low photon count!
Convallaria rhizome (Courtesy: INRA)
25

26. WFM CLSM
High frequencies lost!
Frequency Spectra for Convallaria sample
26

27. λ
d = 0.61
NA
Minimum Radial
Resolution ~ 200nm
... and yet, there is the Diffraction Barrier
27

28. I minimize trade-offs in microscopy ...
SNR
Acquisition
but how?
Speed Resolution
28

29. ... by bridging optics and image Processing
Sample Microscope 3-D Image
• Improvement in
resolution.
• Noise reduction
Diffraction
limit
Aberrations Noise ? ...but requires point-
spread function (PSF)
Inverse h(x).
approach
?
29

30. Point-Spread Function (PSF) h(x)
Separate PSF estimation
Fluorescent Identify Point Blind
PSF Model
Beads Sources estimation
Complexity
Accuracy
30

31. PSF Modeling
Clear Apodized
Pupil Pupil
Defocus Fourier Amplitude
Pupil Transform PSF
Function
Microscope Squared
PSF Magnitude
P. A. Stokseth, “Properties of a defocused
optical system,” J. Opt. Soc. Am. A, vol. 59,
pp. 1314–1321, Oct. 1969.
P. Pankajakshan et al. “Blind Deconvolution for
Confocal Laser Scanning Microscopy,” PhD Thesis.,
31 Dec. 2009.

32. WFM CLSM
−6 −14
−8 −16
−18
−10
−20
z
−12 z
−22
x x
Axial MIP PSF for a 40x/1.4 oil immersion lens.
P. Pankajakshan, et al. Deconvolution and Denoising for Confocal Microscopy.
In F.Cazals and P. Kornprobst, editor, Modeling in Computational Biology and
32 Biomedicine, ch.4, Springer, In Publication, 2012.

33. Determining PSF- Beads as ‘guide stars’
Image Plane
on
ds ce!
an ien
Object Plane
H r
pe
z
xy
ex
Imaging fluorescent beads
33

34. Blind/Myopic Deconvolution
Alternate PSF & Object estimation
PSF constraints Observation Object constraints
Old
Object Update Update New
PSF New PSF Object Object
Old
PSF
T. J. Holmes. Blind deconvolution of quantum-limited
P. Pankajakshan et al. “Blind incoherent imagery: maximum-likelihood approach. J.
Deconvolution for Confocal Laser Opt. Soc. Am. A, 9:1052–1061, July 1992.
Scanning Microscopy,” PhD Thesis., Dec.
2009. B. Zhang et al. “A study of Gaussian approximation of
fluorescence Microscopy PSF models,” SPIE conf., San
34 Jose, Jan. 2006.

35. Convallaria cell from Parenchyma tissue
Observed
Restored
Focussed on the cytoplasmic threads of a cell
P. Pankajakshan, et al. Deconvolution and Denoising for Confocal Microscopy.
In F.Cazals and P. Kornprobst, editor, Modeling in Computational Biology and
35 Biomedicine, ch.4, Springer, In Publication, 2012.

36. Blind Deconvolution of Fluorescent Shell
Specification:
Diameter
15 microns
Thickness
specified:
500-700
nm
2x increase
in resolution!
Observed ThinBlinDe: 40 iter. P. Pankajakshan et al. “Blind
Deconvolution for Confocal Laser
FWHM: 930nm FWHM: 535.58nm
Scanning Microscopy,” PhD Thesis., Dec.
2009.
36

37. Non-Blind vs Blind
Commercial software Proposed ThinBlinDe
(Using PSF model)
37

38. Convallaria cell from Parenchyma tissue
Courtesy: INRA
Commercial ThinBlinDe:
Observed Non-blind Blind
restoration restoration
P. Pankajakshan et al. “Blind Deconvolution for Confocal Laser
Scanning Microscopy,” PhD Thesis., Dec. 2009.
38

39. Thank you
39

40. Contact me ...
praveen.pankaj@gmail.com Praveen Pankajakshan
+33 (0) 628358064
15 rue de l’ésperance
www.bioimageanalysis.org/~praveen
75013 Paris
www.linkedin.com/in/praveenpankaj
pasteur.academia.edu/PraveenPankajakshan
praveenpankaj France
40