Electronic holographic imaging through living human tissue
H. Chen, M. Shih, E. Arons, E. Leith, J. Lopez, D. Dilworth, and P. C. Sun
Electronic holography and a swept-frequency dye laser are used with the first-arriving-Iight method to image an absorbing object through the flesh of a human hand. Holography with living human tissue without the use of high-peak-power lasers is made possible by the high sensitivity of the CCD camera as well as its capability for making a large number of holograms in rapid succession, thus enabling the images to be combined to produce a resultant image with an improved signal-to-noise ratio.
The use of visible light for imaging through or into biological tissue is currently an active area of research.1-7 Some of the most attractive methods utilize the principle of the first-arriving light. A short pulse of light enters the tissue, and because of severe scatter, the light pulse emerging from the opposite side is elongated by a factor of perhaps hundreds or even thousands (Fig. 1). The light that emerges first is scattered the least, or is mostly forward scattered, and therefore produces the best image. Gating methods to reject the later-arriving light enable a moderately good image to be obtained even though the tissue is translucent. Most experiments involving biological tissue have used nonliving tissue, such as chicken meat and excised breast tissue, although living tissue has also been used, such as the transillumination of a living goldfish by Wang et al.8 and of a human hand by Andersson-Engels et al.9 and Gratton et al.10
Holography has been used as a gate.11,12 The scattered light is combined with a reference beam, and the interference pattern is recorded. The reference beam is adjusted in time delay so that it arrives at the recording surface simultaneously with the first-arriving light, thus forming a hologram of only the first-arriving light, the later-arriving light exposing the medium only as an ambient background.
The authors are with the Department of Electrical Engineering and Computer Science, University of Michigan, Ann Arbor, Michigan 48109-2122.
Received 26 July 1993; revised manuscript received 18 February 1994.
With holography, as opposed to many other methods, a short pulse is not required; a cw waveform with a short coherence length is fully equivalent.12 The reference beam is delayed so that it correlates with the first-arriving light, and a fringe pattern is formed with only the first-arriving light, as before. Also, since the exposure process integrates the recorded light, it is not necessary that all wavelength components of the light be generated simultaneously; we can generate one wavelength (or one wavelength band) at a time during the exposure, obtaining a result identical to that obtained when all wavelengths present at once. Thus we generate illumination with the required spectrum and the required coherence function by sweeping a dye laser through a range of wavelengths during the exposure. This is a convenient way of generating the desired coherence function. The holographic method thus has an advantage over many other gating methods, such as the use of streak cameras and Kerr cells, in that a short gate can readily be produced; the gate is just the pulse duration, or equivalently, the coherence length of the illumination.
However, the holographic method has the disadvantage that the fringe contrast will be low because of the ambient background produced by the later-arriving light. The resulting image will then be noisy. In addition, the object and the scattering medium must be stable during the exposure time; otherwise, the fringes will be smeared. This requirement has always been a major problem with holography of human subjects and has been solved by use of pulsed lasers of high peak power, thus permitting short exposure times. For imaging through living tissue the problem is compounded by the high attenuation
Fig. 1. Principle of the first-arriving light for imaging through scattering media.
of the transillumination process as well as by the requirement that the source have a well-controlled autocorrelation function.
We have employed electronic holography, in which the exiting surface of the scattering medium is imaged onto the detecting surface of a CCD camera, a reference beam is introduced, and a hologram is recorded, as in Fig. 2.13 Electronic holography overcomes some of the limitations of conventional holography. The sensitivity of the CCD camera permits measurable exposures to be made with moderate light levels even with an exposure time as short as 5-20 ms, the stability time of typical human living tissue. The great dynamic range (14-18 bits) of the cooled CCD camera permits low-contrast fringes to be recorded. Further, since the holographic reconstruction process in electronic holography is done by computer and can be carried out in as little as a few seconds, and because the CCD camera, as opposed to film, is reusable, a sequence of many holograms can be made in rapid succession, and the images can be added to increase the signal-to-noise ratio. Each exposure is made within a time interval less than the stability time of the tissue, and successive exposures are made at time intervals exceeding the coherence time, thus ensuring that the speckle noise will be uncorrelated between successive holograms; thus averaging over many images will reduce the speckle noise. Other noise, such as photon noise, will similarly be reduced by the averaging process.
Fig. 2. Optical system for electronic holography for imaging through living tissue. The spatial filter reduces the spatial frequency content of the light distribution at the CCD, and the delay structure delays the reference beam by the amount needed to bring it to interference with the first-arriving light.
Fig. 3. Image of wires through transilluminated living human hand.
These capabilities of electronic holography have permitted the formation of first-arriving-light holograms in which the object is an absorbing structure located behind the flesh of a human hand; there is no need for a pulsed laser with attendant high peak power. In the experiment the flesh on the edge of an especially fleshy human hand was drawn out, away from the bone, and squeezed between two glass plates separated by 6 mm. A pair of crossed wires, 2 mm and 1.5 mm in diameter, was placed on the source side of the tissue. As shown in Fig. 2, the exiting surface of the tissue was imaged onto the CCD camera, a reference beam was introduced with a proper delay, and a hologram was recorded. The exposure time had to be long enough to obtain sufficient light with the available 100-mW laser but short enough so that the fringe pattern was not lost because of the tissue instability. A fast mechanical shutter was used, with exposure time down to ~ 1 ms. The optimum exposure time, a compromise between the two above requirements, was found to be ~ 20 ms. Each hologram produced a noisy image (Fig. 3). However, the integration of the images from 25 holograms resulted in a much improved image (Fig. 4) in which the crossed wires were clearly visible.
The swept-frequency dye-laser method of generating a broadband illumination source permits considerable control over the source coherence. Various wavelength sweep ranges were tried, and the optimum was found to be 6 nm, which corresponds to a
Fig. 4. Same as Fig. 3 but with 25 images averaged.
coherence time of ~ 300 fs. The midband of the spectrum was 750 nm, which is close to the optimum wavelength for transillumination of living tissue.
The most significant aspect of this experimental result is the formation of a hologram of (or more precisely, through) a living person with cw light of a rather low level, in contrast with the pulsed high-peak-power light usually used for this purpose. Holographic portraiture, for example, is carried out with a pulse laser that delivers typically 0.1 J/ns, which produces sufficient exposure on the hologram. Typical holographic film requires 500 ergs/cm2 for producing a hologram, and if a fast, nonholographic film, such as tri-x, were used, this requirement reduces to 0.0375 erg/cm2. With a 20-ms exposure time the energy density falling on the recording film would be 2 ergs/cm2. When we consider in addition the large insertion loss produced by transillumination through the tissue, plus the subsequent spatial filtering (needed to reduce the spatial frequency Content of the recorded light distribution so as to match the CCD camera resolution), we find that only ~ 10 -7 of the light incident upon the tissue arrives at the detector. A 100-mW laser would then place only 0.01 (erg/cm2)s on the detector. With a 20-ms exposure time the density of light energy (2 × 10 -4 erg/cm2) would be orders of magnitude too low to make a hologram on even fast photographic film, yet there is adequate light for electronic holography. Success at this light level is made possible first by the superior sensitivity of the cooled CCD detector and second by the capability of the CCD camera for making multiple holograms and summing the images, thus increasing the total exposure and increasing the signal-to-noise ratio.
This work was supported by the National Science Foundation under grants NSF-ECS-9000571 and NSF-STC-PHY-8920108 and by the U.S. Army Research Office under grant DAAL-03-92-G-0230.
References 1. M. A. Duguay and A. T. Mattick, "Ultrahigh speed photogra
phy of picosecond light pulses and echoes," Appl. Opt. 10, 2162-2170 (1971).
2. K. M. Yoo and R. R. Alfano, "Time-resolved coherent and incoherent components of forward light scattering in random media," Opt. Lett. 15, 320-332 (1990).
3. L. Wang, P. P. Ho, C. Liu, G. Zhang, and R. R. Alfano, "Ballistic 2-D imaging through scattering walls using an ultrafast optical Kerr gate," Science 253, 769-771 (1991).
4. J. C. Hebden and R. A. Kruger, "Transillumination imaging performance time of flight imaging system," Med. Phys. 17, 351-356(1990).
5. J. C. Hebden, R. A. Kruger, and K. S. Wong, "Time resolved imaging through a highly scattering medium," Appl. Opt. 30, 788-794(1991).
6. H. J. Gerritsen, "Holography and four wave mixing to see through the skin," in Analog Optical Processing and Computing, H. J. Caulfield, ed., Proc. Soc. Photo-Opt. Instrum. Eng. 519, 128-131 (1984).
7. J. G. Fujimoto, S. De Silversti, E. P. Ippen, R. Margolis, and A. Oseroff, "Femtosecond optical ranging in biological systems," Opt. Lett. 11, 150-152 (1986).
8. L. Wang, Y. Liu, P. P. Ho, and R. R. Alfano, "Ballistic imaging of biological samples using picosecond optical Kerr Gate," in Time-Resolved Spectroscopy and Imaging of Tissues, B. Chance and A. Katzir, eds., Proc. Soc. Photo-Opt. Instrum. Eng. 1431, 97-101 (1991).
9. S. Andersson-Engels, R. Berg, S. Svanberg, and O. Jarlman, "Time resolved transillumination for medical diagnostics," Opt. Lett. 15, 1179-1181 (1990).
10. E. Gratton, W. Mantulin, M. vandeVen, J. Fishkin, M. Maris, and B. Chance, "A novel approach to laser tomography," Bioimaging 1, 40-46 (1993).
11. N. Abramson and K. G. Spears, "Single pulse light-in-flight recording by holography," Appl. Opt. 28, 1834-1841 (1989).
12. K. G. Spears, J. Serafin, N. Abramson, X. Zhu, and H. Bjelkhagen, "Chrono-coherent imaging for medicine," IEEE Trans. Biomed. Eng. 36, 1210-1214 (1989).
13. H. Chen, Y. Chen, D. Dilworth, E. Leith, J. Lopez, and J. Valdmanis, "Two-dimensional imaging through diffusing media using 150-fs gated electronic holography techniques," Opt. Lett. 16,487-489(1991).
