Abstract

Hyperspectral imaging captures rich information in spatial and spectral domains but involves high costs and complex data processing. The use of a set of optical band-pass filters (BPFs) in the acquisition of spectral images is proposed for reducing dimensionality of spectral data while maintaining target detection and/or categorization performance. A set of BPFs that could distinguish ice from surrounding water on various materials (e.g., asphalt), was designed as an example. Relatively high accuracy (90.28%) was achieved with only two BPFs, showing the potential of this approach for accurate target detection with lesser complexity than conventional methods.

© 2012 OSA

Full Article  |  PDF Article

References

  • View by:
  • |
  • |
  • |

  1. N. Gat, “Imaging spectroscopy using tunable filters: a review,” Proc. SPIE 4056, 50–64 (2000).
    [CrossRef]
  2. Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
    [CrossRef] [PubMed]
  3. S. S. Shen and P. E. Lewis, eds., Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIII (Proc SPIE), SPIE-International Society for Optical Engine (2007).
  4. K. Sasaki, S. Kawata, and S. Minami, “Optimal wavelength selection for quantitative analysis,” Appl. Spectrosc. 40(2), 185–190 (1986).
    [CrossRef]
  5. J. M. Brenchley, U. Horchner, and J. H. Kalivas, “Wavelength selection characterization for NIR spectra,” Appl. Spectrosc. 51(5), 689–699 (1997).
    [CrossRef]
  6. S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
    [CrossRef]
  7. B. Smith, “Wavelength selection and optimization of pattern recognition methods using the genetic algorithm,” Anal. Chim. Acta 423(2), 167–177 (2000).
    [CrossRef]
  8. R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
    [CrossRef] [PubMed]
  9. P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
    [CrossRef]
  10. A. Troiano, E. Pasero, and L. Mesin, “New system for detecting road ice formation,” IEEE Trans. Instrumentat. Meas. 60(3), 1091–1101 (2011).
    [CrossRef]
  11. R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
    [CrossRef]
  12. C. E. Bassey and G. R. Simpson, “Aircraft ice detection using time domain reflectometry with coplanar sensors,” in 2007 IEEE Aerospace Conference (IEEE, 2007), pp. 1–6.
  13. M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
    [CrossRef]
  14. R. Pearce, “Plant freezing and damage,” Ann. Bot. (Lond.) 87(4), 417–424 (2001).
    [CrossRef]
  15. D. Gregoris, S. Yu, and F. Teti, “Multispectral imaging of ice” in Canadian Conference on Electrical and Computer Engineering (2004), vol. 4, pp. 2051–2056.
  16. E. Angelopoulou, R. Molana, and K. Danilidis, “Multispectral skin color modelling,” in Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2001. CVPR 2001 (IEEE/COMSOC, 2001), vol. 2, pp. II-635–II-642.
  17. S. J. Preece and E. Claridge, “Spectral filter optimization for the recovery of parameters which describe human skin,” IEEE Trans. Pattern Anal. Mach. Intell. 26(7), 913–922 (2004).
    [CrossRef] [PubMed]
  18. J. Karlholm and I. Renhorn, “Wavelength band selection method for multispectral target detection,” Appl. Opt. 41(32), 6786–6795 (2002).
    [CrossRef] [PubMed]
  19. O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
    [CrossRef]
  20. J. Minet, J. Taboury, F. Goudail, M. Péalat, N. Roux, J. Lonnoy, and Y. Ferrec, “Influence of band selection and target estimation error on the performance of the matched filter in hyperspectral imaging,” Appl. Opt. 50(22), 4276–4285 (2011).
    [CrossRef] [PubMed]
  21. K. Nishino, M. Nakamura, M. Matsumoto, O. Tanno, and S. Nakauchi, “Optical filter highlighting spectral features part II: quantitative measurements of cosmetic foundation and assessment of their spatial distributions under realistic facial conditions,” Opt. Express 19(7), 6031–6041 (2011).
    [CrossRef] [PubMed]
  22. K. Nishino, M. Nakamura, M. Matsumoto, O. Tanno, and S. Nakauchi, “Optical filter for highlighting spectral features part I: design and development of the filter for discrimination of human skin with and without an application of cosmetic foundation,” Opt. Express 19(7), 6020–6030 (2011).
    [CrossRef] [PubMed]
  23. R. Fisher, “The use of multiple measurements in taxonomic problems,” Ann. Eugen. 7(2), 179–188 (1936).
    [CrossRef]

2011 (4)

2006 (2)

Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
[CrossRef] [PubMed]

M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
[CrossRef]

2004 (2)

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

S. J. Preece and E. Claridge, “Spectral filter optimization for the recovery of parameters which describe human skin,” IEEE Trans. Pattern Anal. Mach. Intell. 26(7), 913–922 (2004).
[CrossRef] [PubMed]

2003 (1)

O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
[CrossRef]

2002 (1)

2001 (1)

R. Pearce, “Plant freezing and damage,” Ann. Bot. (Lond.) 87(4), 417–424 (2001).
[CrossRef]

2000 (2)

B. Smith, “Wavelength selection and optimization of pattern recognition methods using the genetic algorithm,” Anal. Chim. Acta 423(2), 167–177 (2000).
[CrossRef]

N. Gat, “Imaging spectroscopy using tunable filters: a review,” Proc. SPIE 4056, 50–64 (2000).
[CrossRef]

1999 (1)

R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
[CrossRef] [PubMed]

1998 (1)

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

1997 (2)

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

J. M. Brenchley, U. Horchner, and J. H. Kalivas, “Wavelength selection characterization for NIR spectra,” Appl. Spectrosc. 51(5), 689–699 (1997).
[CrossRef]

1986 (1)

1936 (1)

R. Fisher, “The use of multiple measurements in taxonomic problems,” Ann. Eugen. 7(2), 179–188 (1936).
[CrossRef]

Brenchley, J. M.

Chan, D. E.

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

Chen, Y.-R.

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

Claridge, E.

S. J. Preece and E. Claridge, “Spectral filter optimization for the recovery of parameters which describe human skin,” IEEE Trans. Pattern Anal. Mach. Intell. 26(7), 913–922 (2004).
[CrossRef] [PubMed]

Destain, M.-F.

O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
[CrossRef]

Edwards, D. A.

R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
[CrossRef] [PubMed]

Feo, A.

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Ferrec, Y.

Fisher, R.

R. Fisher, “The use of multiple measurements in taxonomic problems,” Ann. Eugen. 7(2), 179–188 (1936).
[CrossRef]

Garini, Y.

Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
[CrossRef] [PubMed]

Gat, N.

N. Gat, “Imaging spectroscopy using tunable filters: a review,” Proc. SPIE 4056, 50–64 (2000).
[CrossRef]

Golia, C.

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Goudail, F.

Homola, M.

M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
[CrossRef]

Horchner, U.

Jordan, R. B.

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

Kalivas, J. H.

Karlholm, J.

Kawata, S.

Kim, M. S.

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

Kind, R.

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Kleynen, O.

O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
[CrossRef]

Künnemeyer, R.

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

Leemans, V.

O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
[CrossRef]

Lonnoy, J.

Matsumoto, M.

McNamara, G.

Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
[CrossRef] [PubMed]

Mehl, P.

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

Mesin, L.

A. Troiano, E. Pasero, and L. Mesin, “New system for detecting road ice formation,” IEEE Trans. Instrumentat. Meas. 60(3), 1091–1101 (2011).
[CrossRef]

Minami, S.

Minet, J.

Nakamura, M.

Nakauchi, S.

Nicklasson, P.

M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
[CrossRef]

Nishino, K.

Ornberg, R. L.

R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
[CrossRef] [PubMed]

Osborne, S. D.

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

Pasero, E.

A. Troiano, E. Pasero, and L. Mesin, “New system for detecting road ice formation,” IEEE Trans. Instrumentat. Meas. 60(3), 1091–1101 (2011).
[CrossRef]

Péalat, M.

Pearce, R.

R. Pearce, “Plant freezing and damage,” Ann. Bot. (Lond.) 87(4), 417–424 (2001).
[CrossRef]

Potapczuk, M.

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Preece, S. J.

S. J. Preece and E. Claridge, “Spectral filter optimization for the recovery of parameters which describe human skin,” IEEE Trans. Pattern Anal. Mach. Intell. 26(7), 913–922 (2004).
[CrossRef] [PubMed]

Renhorn, I.

Roux, N.

Sasaki, K.

Shah, A. D.

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Smith, B.

B. Smith, “Wavelength selection and optimization of pattern recognition methods using the genetic algorithm,” Anal. Chim. Acta 423(2), 167–177 (2000).
[CrossRef]

Sundsbø, P. A.

M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
[CrossRef]

Taboury, J.

Tanno, O.

Troiano, A.

A. Troiano, E. Pasero, and L. Mesin, “New system for detecting road ice formation,” IEEE Trans. Instrumentat. Meas. 60(3), 1091–1101 (2011).
[CrossRef]

Woerner, B. M.

R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
[CrossRef] [PubMed]

Young, I. T.

Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
[CrossRef] [PubMed]

Anal. Chim. Acta (1)

B. Smith, “Wavelength selection and optimization of pattern recognition methods using the genetic algorithm,” Anal. Chim. Acta 423(2), 167–177 (2000).
[CrossRef]

Analyst (Lond.) (1)

S. D. Osborne, R. Künnemeyer, S. D. Osborne, and R. B. Jordan, “Method of wavelength selection for Partial Least Squares,” Analyst (Lond.) 122(12), 1531–1537 (1997).
[CrossRef]

Ann. Bot. (Lond.) (1)

R. Pearce, “Plant freezing and damage,” Ann. Bot. (Lond.) 87(4), 417–424 (2001).
[CrossRef]

Ann. Eugen. (1)

R. Fisher, “The use of multiple measurements in taxonomic problems,” Ann. Eugen. 7(2), 179–188 (1936).
[CrossRef]

Appl. Opt. (2)

Appl. Spectrosc. (2)

Cold Reg. Sci. Technol. (1)

M. Homola, P. Nicklasson, and P. A. Sundsbø, “Ice sensors for wind turbines,” Cold Reg. Sci. Technol. 46(2), 125–131 (2006).
[CrossRef]

Cytometry A (1)

Y. Garini, I. T. Young, and G. McNamara, “Spectral imaging: principles and applications,” Cytometry A 69A(8), 735–747 (2006).
[CrossRef] [PubMed]

IEEE Trans. Instrumentat. Meas. (1)

A. Troiano, E. Pasero, and L. Mesin, “New system for detecting road ice formation,” IEEE Trans. Instrumentat. Meas. 60(3), 1091–1101 (2011).
[CrossRef]

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

S. J. Preece and E. Claridge, “Spectral filter optimization for the recovery of parameters which describe human skin,” IEEE Trans. Pattern Anal. Mach. Intell. 26(7), 913–922 (2004).
[CrossRef] [PubMed]

J. Food Eng. (1)

P. Mehl, Y.-R. Chen, M. S. Kim, and D. E. Chan, “Development of hyperspectral imaging technique for the detection of apple surface defects and contaminations,” J. Food Eng. 61(1), 67–81 (2004).
[CrossRef]

J. Histochem. Cytochem. (1)

R. L. Ornberg, B. M. Woerner, and D. A. Edwards, “Analysis of stained objects in histological sections by spectral imaging and differential absorption,” J. Histochem. Cytochem. 47(10), 1307–1313 (1999).
[CrossRef] [PubMed]

Opt. Express (2)

Postharvest Biol. Technol. (1)

O. Kleynen, V. Leemans, and M.-F. Destain, “Selection of the most efficient wavelength bands for ‘Jonagold’ apple sorting,” Postharvest Biol. Technol. 30(3), 221–232 (2003).
[CrossRef]

Proc. SPIE (1)

N. Gat, “Imaging spectroscopy using tunable filters: a review,” Proc. SPIE 4056, 50–64 (2000).
[CrossRef]

Prog. Aerosp. Sci. (1)

R. Kind, M. Potapczuk, A. Feo, C. Golia, and A. D. Shah, “Experimental and computational simulation of in-flight icing phenomena,” Prog. Aerosp. Sci. 34(5-6), 257–345 (1998).
[CrossRef]

Other (4)

C. E. Bassey and G. R. Simpson, “Aircraft ice detection using time domain reflectometry with coplanar sensors,” in 2007 IEEE Aerospace Conference (IEEE, 2007), pp. 1–6.

D. Gregoris, S. Yu, and F. Teti, “Multispectral imaging of ice” in Canadian Conference on Electrical and Computer Engineering (2004), vol. 4, pp. 2051–2056.

E. Angelopoulou, R. Molana, and K. Danilidis, “Multispectral skin color modelling,” in Proceedings of the 2001 IEEE Computer Society Conference on Computer Vision and Pattern Recognition, 2001. CVPR 2001 (IEEE/COMSOC, 2001), vol. 2, pp. II-635–II-642.

S. S. Shen and P. E. Lewis, eds., Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XIII (Proc SPIE), SPIE-International Society for Optical Engine (2007).

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 (14)

Fig. 1
Fig. 1

Basic concept used to reduce the dimensionality of a given spectral data set by selecting wavelength bands with band-pass filtering.

Fig. 2
Fig. 2

Schematic diagram of experiment setup (left) and measured samples (right). The sample in the near side of the image is the water sample and the one behind it is the ice sample. After the ice sample melted, it was used as a water sample.

Fig. 3
Fig. 3

Radiance spectra of ice (solid red line) and water (dashed blue line) on asphalt. The thin lines show the original spectra and the bold lines show the average spectra for ice and water.

Fig. 4
Fig. 4

Flowchart of the multiple selection method, which consists of a global search process (left column) and a local optimization process (right column).

Fig. 5
Fig. 5

Optimization results: the two optimal designed bands and the two optimal wavelengths; superimposed mean and individual spectra shown in Fig. 3. The Fisher criteria J in both cases are displayed at the top of the figure.

Fig. 6
Fig. 6

Comparison of the Fisher criteria J as computed by conventional spectral analysis techniques (MCS + PCA + LDA) for various wavelength regions, for two wavelengths (Two WL), and for the proposed two BPFs (Two bands).

Fig. 7
Fig. 7

Distributions of the discriminant score defined by Eq. (6) for the two wavelengths (Two WL), the proposed two BPFs (Two bands) and the multivariate cases (MSC + PCA + LDA) for 900–2500 and 900–2370 nm.

Fig. 8
Fig. 8

The Fisher criteria J of the number of filters selected using the multiple selection method.

Fig. 9
Fig. 9

Theoretical and realized spectral transmittances of tested filters. The computed Fisher criterion is displayed at the top of the figure.

Fig. 10
Fig. 10

Measurement setup: a near-infrared camera XSVA XS FPA-1.7–320 (Xenics) with installed developed filters was used to take the spectral images. Two sets of halogen lamps (300 W each) illuminated the samples from both sides. The camera was set just above the sample.

Fig. 11
Fig. 11

Visualization results: the left image shows the RGB color image taken with a digital camera, and the right image shows the detection result computed from the images taken with the NIR camera with the installed optimized filters. The top sample had a 1-cm-thick ice layer and the bottom one had a water layer. The pixels classified as ice were colored red and the areas classified as water were colored blue.

Fig. 12
Fig. 12

Visualization results over time, which clearly show the area classified as ice (the sample in the top row) decreasing with time.

Fig. 13
Fig. 13

Visualization results for other materials: ice on leaf (93.56%) and glass (98.99%). Colors are shown in the same manner as in Figs. 11 and 12.

Fig. 14
Fig. 14

Evaluation index calculated using the multiple selection method and the transmittance functions of the selected filters (N = 2). (a), (b) when the first filter is selected (k = 2); and (c), (d) when the second filter is selected (k = 2). The horizontal axis shows the bandwidth and the vertical axis shows the center wavelength of BPFs. The gray level corresponds to the average Fisher criterion J of the given filter parameters. The white arrow in (a) points to the wavelength selected using the two-wavelength method, and the arrow in (c) points to the parameter of the second filter selected using the multiple selection method.

Tables (1)

Tables Icon

Table 1 Measurement conditions

Equations (7)

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

O i = T i (λ)S(λ)I(λ)dλ ,
T i (λ)={ 1 for λ L i λ λ H i 0 for λ< λ L i , λ> λ H i ,
J(w)= w t Σ BWN w w t Σ WTN w ,
Σ WTN = Σ A + Σ B , Σ A,B = OA,B ( O A,B O ¯ A,B ) ( O A,B O ¯ A,B ) t t
w*= argmax w J(w)= Σ WTN 1 ( O ¯ A O ¯ B ).
f(O)= 1 2 { (O μ A ) t Σ 1 (O μ A ) (O μ B ) t Σ 1 (O μ B ) } = { Σ 1 ( μ A μ B ) } t O 1 2 ( μ A + μ B ) t Σ 1 ( μ A μ B ).
J({ λ L i , λ H i })= w * t Σ BWN w* w * t Σ WTN w* = { f( O A ) ¯ - f( O B ) ¯ } 2 var[f( O A ]+var[f( O B ] ,

Metrics