Multivariate image analysis 411
11. Indahl UG, Næs T. Evaluation of alternative spectral fea-
ture extraction methods of textual images for multivari-
ate modeling. J. Chemometrics 1998; 12: 261–278.
12. Gouti N, van Espen P, Feinberg MH. Quantitative recon-
struction of objects from spatially correlated image
sequences. Chemometrics Intell. Lab. Syst. 1999; 47: 21–31.
13. Artyushkova K, Fulghum JE. Multivariate image analy-
sis methods applied to XPS imaging data sets. Surf. Inter-
face Anal. 2002; 33: 185–195.
14. Courcoux P, Devaux M-F, Bouchet B. Simultaneous
decomposition of multivariate images using three-way
data analysis. Application to the comparison of cereal
grains by confocal laser scanning microscopy. Chemo-
metrics Intell. Lab. Syst. 2002; 62: 103–113.
A general approach for the analysis of time-resolved
image data has been presented. The multivariate modeling
of movies has been successfully used to identify features
with different dynamic profiles. It can easily be envisaged
that by fitting kinetic models to the profiles found, as has
already been done within standard spectrometry [44,45],
imaging could be used as a means of identifying chemical
constants such as diffusion rates. Application of this meth-
odology is not limited to video imaging, but is relevant to
any form of chemical imaging in which measurement at
regular time intervals is possible.
As has already been documented, one of the major pro-
blems in image analysis is the computational burden levied
by the huge data arrays being produced. Whilst multivariate
modeling techniques such as the subspace decompositions
described here can play an important role, it is also useful to
consider how image analysis and processing techniques—of
which the histogram and mapping tools illustrated here are
only a small part—can be used to extract the relevant
chemical information from the data. One challenge in che-
mical imaging will be the combination of the multivariate
statistical analysis methods already proving successful in
chemistry with the vast range of image analysis tools already
available in the image analysis sciences.
´
15. Lage EM. Fluxo de CO2 na interface agua–ar. PhD Thesis,
Universidade Estadual de Campinas (UNICAMP), 2002.
16. Esbensen KH, Wold S, Geladi P. Relationships between
higher-order data array configurations and problem for-
mulations in multivariate data analysis. J. Chemometrics
1988; 3: 33–48.
17. Geladi P, Grahn H. Multivariate Image Analysis. Wiley:
Chichester, 1996.
18. Wold S, Geladi P, Esbensen K, Ohman J. Multi-way prin-
cipal components- and PLS-analysis. J. Chemometrics
1987; 1: 41–56.
19. Geladi P, Grahn H, Esbensen K, Bengtsson E. Image ana-
lysis in chemistry. II. Multivariate image analysis. Trends
Anal. Chem. 1992; 11: 121–130.
20. Harshman RA, Lundy ME. PARAFAC—parallel factor
analysis. Comput. Statist. Data Anal. 1997; 18: 39–72.
21. Bro R. PARAFAC. Tutorial and applications. Chemo-
metrics Intell. Lab. Syst. 1997; 38: 149–171.
¨
Acknowledgements
The authors gratefully acknowledge financial support
from the State of Sa˜o Paulo Research Foundation (FAPESP)
and the Brazilian National Research Council (CNPq).
22. Smilde AK. 3-way analyses—problems and prospects.
Chemometrics Intell. Lab. Syst. 1992; 15: 143–157.
23. Gurden SP, Westerhuis JA, Bro R, Smilde AK. A compar-
ison of multiway regression and scaling methods. Che-
mometrics Intell. Lab. Syst. 2001; 59: 121–136.
´ ´
24. Sarmiento JL, Murnane R, le Quere C. Air–sea CO2 trans-
REFERENCES
fer and the carbon budget of the North Atlantic. Philos.
Trans. R. Soc. Lond. B 1995; 348: 211–219.
25. Sarmiento JL, Toggweiler JR, Najjar R. Ocean carbon-
cycle dynamics and atmospheric pCO2. Philos. Trans. R.
Soc. Lond. A 1988; 325: 3–21.
1. Bhargava R, Levin IR. Fourier transform infrared ima-
ging: theory and practice. Anal. Chem. 2001; 73: 5157–
5167.
2. Tran CD. Development and analytical applications of
multispectral imaging techniques: an overview. Fresen-
tus J. Anal. Chem. 2001; 369: 313–319.
3. Tran CD. Visualising chemical composition and
reaction kinetics by the near infrared multispectral
imaging technique. J. Near Infrared Spectrosc. 2000; 8:
87–99.
4. Macpherson JV, Jones CE, Barker AL, Unwin PR. Elec-
trochemical imaging of diffusion through single nanos-
cale pores. Anal. Chem. 2002; 74: 1841–1848.
26. Jahne B, Haussecker H. Air–water gas exchange. Ann.
¨
Rev. Fluid Mech. 1998; 30: 443–468.
27. Phillips LF. Experimental demonstration of coupling of
heat and matter fluxes at a gas–water interface. J. Geo-
phys. Res. 1994; 99: 18577–18584.
28. Wanninkhof R, Asher W, Weppernig R, Chen H, Schlos-
ser P, Langdon C, Sambrotto R. Gas transfer experiment
on Georges Bank using two volatile deliberate tracers. J.
Geophys. Res. 1993; 98: 20237–20248.
29. Erickson III DJ. A stability dependent theory for air–sea
gas exchange. J. Geophys. Res. 1993; 98: 8471–8488.
30. Jahne B, Munnich KO, Bosinger R, Dutzi A, Huber W,
5. Bharati MH, MacGregor JF. Multivariate image analysis
for real-time process monitoring and control. Ind. Eng.
Chem. Res. 1998; 37: 4715–4724.
¨
¨
¨
Libner P. On the parameters influencing air–water gas
exchange. J. Geophys. Res. 1987; 92: 1937–1949.
31. Wanninkhof R. Relationship between wind speed and
gas exchange over the ocean. J. Geophys. Res. 1992; 97:
7373–7382.
32. Phillips LF. Steady-state thermodynamics of transfer
through a gas–liquid interface, treated as a limiting case
of thermo-osmosis. Chem. Phys. Lett. 1994; 228: 533–538.
33. Andersson CA, Bro R. Chemometrics Intell. Lab. Syst. 2000;
52: 1–4.
34. The N-way Toolbox for MATLAB [Online]. Available:
February 2003].
35. Vogt F, Tacke M. Fast principal component analysis of
large data sets. Chemometrics Intell. Lab. Syst. 2001; 59:
1–18.
6. van Espen P, Janssens G, Vanhoolst W, Geladi P. Ima-
ging and image processing in analytical chemistry. Ana-
lusis 1992; 20: 81–90.
7. Geladi P, Isaksson H, Lindqvist L, Wold S, Esbensen K.
Principal component analysis of multivariate images.
Chemometrics Intell. Lab. Syst. 1989; 5: 209–220.
8. Esbensen K, Geladi P. Strategy of multivariate image
analysis (MIA). Chemometrics Intell. Lab. Syst. 1989; 7:
67–86.
9. Lied TT, Geladi P, Esbensen KH. Multivariate image
regression (MIR): implementation of image PLSR—first
forays. J. Chemometrics 2000; 14: 585–598.
10. Huang J, Esbensen KH. Applications of angle measure
technique (AMT) in image analysis. Part I. A new meth-
odology for in situ powder characterization. Chemo-
metrics Intell. Lab. Syst. 2000; 54: 1–19.
Copyright # 2003 John Wiley & Sons, Ltd.
J. Chemometrics 2003; 17: 400–412