Development of Tunable Fabry-Perot Spectral Camera and Light Source for Medical Applications M. Kaarre a, S. Kivi a, P.E. Panouillot a, H. Saari a, J. Mäkynen a, I. Sorri b and M. Juuti a a VTT Technical research centre of Finland, P.O. Box 1199, FI-70211 Kuopio, Finland b Department of Ophthalmology, Kuopio University Hospital, Kuopio, Finland Abstract. VTT has developed a fast, tunable Fabry-Perot (FP) filter component and applied it in making small, lightweight spectral cameras and light sources. One application field where this novel technology is now tested is medical field. A demonstrator has been made to test the applicability of FP based spectral filtering in the imaging of retina in visible light wavelength area. Keywords: Fundus microscope, retina imaging, Fabry-Perot filter. PACS: 42.79.Ci INTRODUCTION Spectral imaging of the human retina is a useful tool for detecting different types of abnormalities caused by diabetes, glaucoma or age related macular degeneration. Different features, such as microaneurysms or fibrosis are best seen at different colors, i.e. different wavelengths: therefore a multispectral image is needed in order to distinguish between the various details in the retina. The spectral characteristics of typical retinal features have been identified in previous studies [1-5]. Traditionally ophthalmic screening has been done with RGB fundus photography, but the spectral resolution of RGB imaging is very limited and does not provide in-depth information on the nature of structures seen in the image. Several types of multispectral fundus imaging systems have been developed, some of which use broad spectrum illumination and a set of narrow band filters in front of the camera [2,3], while others have applied narrow band filtering to the illumination [1,5]. The latter approach is highly beneficial because the eye can only be exposed to a limited amount of light: now the illuminating intensity at a narrow-band wavelength can be much higher, which in turn enables shorter exposure times of the camera. This is important because of rapid involuntary eye movements: the images get fuzzy with long exposure times. Wavelength filtering in fundus imaging has conventionally been implemented by filter wheels [1], or more recently by Liquid Crystal Tunable Filters (LCTF) [2,5], but also diffractive optical elements [6] and digital micromirror devices [7] have been used. VTT has developed a novel, tunable Fabry-Perot Interferometer module [8,9] which has already been implemented in a handheld stand-alone spectral camera [10,11]. In this work we have applied the tunable Fabry-Perot filter in a controllable illuminator, which is coupled with an ophthalmic fundus camera. FABRY-PEROT TUNABLE LIGHT SOURCE The optical concept of the new Piezo actuated Fabry-Perot tunable light source is shown in Figure 1. The preliminary broadband light source is a white LED, CREE X-Lamp X-GP12. The maximum DC current of this LED is 1.5 A and forward voltage 3.25 V, i.e. the electrical power is 4.9 W. The chip size of this LED is 1.2 mm x 1.2 mm, which enables forming a highly collimated light beam with a single aspheric collimator lens. The collimated light beam propagates via a long pass filter through the Piezo-Actuated Fabry-Perot Interferometer which transmits the spectral band defined by the air gap between the FPI mirrors. The transmitted light is focused into a fiber bundle that is connected to the Canon CR5-45NM fundus camera.
FIGURE 1. Optical concept of the PFPI tunable light source Piezo-Actuated Fabry-Perot Interferometer (PFPI) VTT started to develop Piezo actuated Fabry-Perot Interferometer (PFPI) modules in 2007. Several different FPI constructions have been tried, and manufactured modules have been used in several applications [10,11]. The wavelength selective component in the LED Light source is the Fabry-Perot interferometer which consists of two semi-transparent mirrors placed face to face. The two mirrors create an optical resonator. Constructive interference allows certain wavelengths to be transmitted through the interferometer whereas other wavelengths are reflected. Figures 2 and 3 describe details of the PFPI module. The parallelism and the distance between the interferometer mirrors need to be controlled with high accuracy. This is achieved with three closed loop control channels. Each piezoelectric actuator has a closely positioned capacitive measurement point to determine the mirror separation. Each channel is controlled with nanometer stability and repeatability to obtain the desired parallelism and air gap between the mirrors. FIGURE 2. Structure of the Piezo-actuated Fabry-Perot Interferometer (PFPI) Module.
FIGURE 3. In the tunable light source the PFPI module is covered with a light shield. Figure 4 shows the transmission of the Fabry-Perot Interferometer with three different air gap values of 250 nm, 280 nm and 310 nm. Each of these example air gap values allows two transmission bands to go through the FPI. The emission spectrum of the white LED is limited to the range 420 770 nm. Therefore the transmission bands above 800 nm in the example of Figure 4 can be omitted. FIGURE 4. Transmission of the Fabry-Perot interferometer with air gap values of 250 nm, 280 nm and 310 nm. Control logic of the PFPI Tunable light source A block diagram of the PFPI tunable light source control electronics is shown in Figure 5. In the design we will limit the maximum on time to 3 ms. This gives a duty cycle of 0.3 with 100 Hz frequency. The LPC1343 ARM cortex M3 microcontroller was selected for the application. It provides the USB-control interface for the measurement PC so that the system is fully controllable from the PC. A Labview control program has been developed for the controlling.
FIGURE 5. Block diagram of the PFPI control system. FIGURE 6. Photograph of the PFPI tunable light source.
Characterization of the PFPI Tunable Light Source The 10.5 mm aperture Piezoactuated PFPI component can be used in the wavelength range 400 1000 nm. The measured spectral resolution is 5 20 nm @ FWHM in the usable air gap range 300 3000 nm. The light source can be used with fiber optic bundles with diameter in the range 0.5 15 mm. The light source can provide spectral photon fluxes in the range 1014 1015 photons/s/mm2 at the selected spectral band of 10 20 nm @ FWHM. FIGURE 7. Spectral photon flux measurement result for the PFPI tunable light source. RETINA IMAGING SETUP We used a Canon CR5-45NM fundus camera, which was modified for spectral imaging. The original light source was removed and replaced by the new Fabry-Perot tunable light source, from which the light was guided into the fundus camera by an optical fiber bundle (Figure 8). The volunteers pupils were dilated with topical tropicamidephenylephrine (0.8%+5%) drops about 20 minutes before imaging. A Labsphere white reflectance standard was used for taking white reference images for image analysis purposes. FIGURE 8. Commercial fundus camera modified with the Fabry-Perot tunable wavelength light source. Artificial eye shown in place of a patient s eye.
Spectral Fundus Imaging with FP Light Source The goal of the fundus imaging related to Fabry-Perot light source was to determine the quality of the spectral images obtained. The images are all from healthy eyes as this study didn t include research on systematic deceases. One studied parameter was the power of the light source related to the question what is the minimum integration time of the camera to produce quality images with the source? This is an important question because of the sensory movements of the eye easily blurs the image when using long integration times. The second studied question is related to the spatial and spectral resolution. Can we observe the small veins and other details? There are also known applications of spectral imaging like the ability of imaging deeper layers of retina with longer wavelengths. Figure 9 shows a set of spectral images from one of the volunteer s retina. The images clearly show that with shorter wavelengths the surface of the retina is imaged and when wavelength increases, the imaging depth also increases. Finally with red light, the vascular choroid layer behind the retina becomes visible. The visibility of the choroid layer is related to the amount of macular pigment. In the spectral fundus images of the volunteers we observed a substantial variation in the amount of macular pigment (MP), which is normal among healthy people. In fact, spectral imaging has been used for quantifying the MP it has been shown that the MP density correlates with various health related parameters in humans [13], but a clinically acceptable and easily accessible method to measure the MP density in vivo has been lacking [14]. Our results show that the Fabry-Perot light source could also be well suitable for this application. FIGURE 9. Images from healthy eye by using Fabry-Perot light source. The images clearly show the ability to see the vascular layer of the eye (Choroid) with longer wavelengths.
Figure 10 shows a zoomed image from the optic nerve head area. This image shows that the spatial resolution of the imaging is good as the small veins are visible and the eye doesn t move very easily during 150 ms integration time of the camera. However, compared to the 450 ms integration time of Figure 9 images, the image is a bit grainy. FIGURE 10. Zoomed image showing the details from Optic nerve head area. Integration time was 150 ms. CONCLUSIONS The piezo-actuated Fabry-Perot camera and light source has been developed and tested in medical applications. This paper reviewed the use of FP light source in retina imaging. According to ten healthy volunteer s retina imaging, the quality of the images is sufficient but more light would be needed to make them less grainy. The complete evaluation of the medical applicability of the system would require tests with real diabetes, glaucoma and AMD patients. These are planned to be done as a next step of the system testing. ACKNOWLEDGMENTS The authors would like to thank the InFotonics Center of University of Eastern Finland for equipment and discussion. This work has been supported by Tekes, the Finnish Funding Agency for Technology and Innovation, and European Regional Development Fund. REFERENCES 1. F. C. Delori, E. S. Gragoudas, R. Francisco and R. C. Pruett, Monochromatic ophthalmoscopy and fundus photography, Arch. Ophthalmol. 95, 861-868 (1977) 2 B. Styles, A. Calcagni, E. Claridge, F. Orihuela-Espina and J. M. Gibson, Quantitative analysis of multi-spectral fundus images, Medical Image Analysis 10, 578-597 (2006) 3 J. C. Ramella-Roman, S. A. Mathews, H. Kandimalla, A. Nabili, D. D. Duncan, S. A. D Anna, S. M. Shah and Q. D. Nguyen, Measurement of oxygen saturation in the retina with a spectroscopic sensitive multi aperture camera, Optics Express 16, 6170-6182 (2008) 4 P. Fält, J. Hiltunen, M. Hauta-Kasari, I. Sorri, V. Kalesnykiene, and H. Uusitalo, Extending diabetic retinopathy imaging from color to spectra, Proceedings of Image Analysis, 16th Scandinavian Conference, Lecture Notes in Computer Science (LNCS) (Springer, Berlin, 2009) Vol. 5575, pp. 149 158 5 P. Fält, J. Hiltunen, M. Hauta-Kasari, I. Sorri, V. Kalesnykiene, J. Pietilä and H. Uusitalo, Spectral imaging of the human retina and computationally determined optimal illuminants for diabetic retinopathy lesion detection, J. Imaging Sci. Technol. 55, 1-1 1-10 (2011) 6 W. R. Johnson, D. W. Wilson, W. Fink, M. Humayun and G. Bearman, Snapshot hyperspectral imaging in ophthalmology, Journal of Biomedical Optics 12, 014036-1 014036-7 (2007)
7 R. P. Francis, K. J. Zuzak and R. Ufret-Vincenty, Hyperspectral retinal imaging with a spectrally tunable light source, Proc. of SPIE Vol. 7932, 793206-1 793206-8 (2011) 8 H. Saari, Spectrometer and interferometric method, Patent FI119830B (2009) 9 H. Saari, Spectrometer and interferometric method, Pat. appl. WO2007/135244 (2007) 10 H. Saari, V.-V. Aallos, C. Holmlund, J. Malinen and J. Mäkynen, Handheld hyperspectral imager, Proc. SPIE 7680, 76800D, (2010) 11 H. Saari, V.-V. Aallos, A. Akujärvi, T. Antila, C. Holmlund, U. Kantojärvi, J. Mäkynen and J. Ollila, Novel miniaturized hyperspectral sensor for UAV and space applications, Proc. SPIE 7474, 74741M (2009) 12 XLamp X-PG data sheet down loaded 14.4.2012 http://www.cree.com/~/media/files/cree/led%20components%20and%20modules/xlamp/data%20and%20binning/xla mpxpg.pdf 13 P. S. Bernstein, F. C. Delori, S. Richer, F. J.M. van Kuijk and A. J. Wenzel, The value of measurement of macular carotenoid pigment optical densities and distributions in age-related macular degeneration and other retinal disorders, Vision Research 50, 716 728 (2010) 14 A. A. Fawzi, N. Lee, J. H. Acton, A. F. Laine and R. T. Smith, Recovery of macular pigment spectrum in vivo using hyperspectral image analysis, Journal of Biomedical Optics 16, 106008-1 106008-9 (2011)