A technique has been developed to image a fluorescent object hiding behind a light-scattering screen without the need for a detector behind the screen. The approach could find applications in imaging biological tissue. See Letter p.232
A golfer faced with the problem of hitting a ball out of the woods after an errant shot sometimes makes a brave choice: she aims straight for the trees, swinging the club as hard as possible in the hope that the ball will bounce off the trees and miraculously emerge from the woods. On page 232 of this issue, Bertolotti et al.1 describe a technique for imaging objects through light-scattering media, such as fog and human tissue, that overcomes a challenge that is in some ways similar to this one.
Consider light from a torch passing through a human hand. Information about the shapes of the bones, or even the cells, that make up the hand is thought to be encrypted in this transmitted light (Fig. 1), but a simple device such as a lens cannot be used to image the hand's interior. Numerous attempts have been made to retrieve the shapes of objects that hide behind or are within media that transmit and scatter light. Some of the photons that travel through a light-scattering medium do so without interacting with any of the medium's constituent matter. Such 'ballistic' photons exit the medium a little earlier than their non-ballistic counterparts, which bounce off the matter as they pass through the scattering medium. If the ballistic photons alone are captured in a detector, the blurring effects of scattering can be avoided2. However, for strongly scattering media such as a human hand, ballistic light can propagate only short distances (about 1 millimetre in human tissue) without scattering. Therefore, image quality rapidly degrades as we attempt to see deeper into the tissue3,4.
Another approach for seeing through a scattering medium is to use a technique called phase conjugation. In this method, the paths of all the photons are reversed and, as the photons travel backwards, the scattering that occurred over the course of their forward paths is undone5,6,7. We can understand how this works by returning to the golf analogy. To hit the ball out of the woods, the player would need to have memorized the exact direction in which the ball was travelling when it hit the ground. If she could then hit the ball in precisely the same direction but backwards, the ball would retrace its path and exit the woods. Unfortunately, the ball would end up at the location from which the errant shot was made, and not near the hole as the player would wish. Phase conjugation has a similar problem. A double pass through the same scattering medium gives a well-focused image of the object. However, this image forms right next to the object hiding behind the scattering medium, and thus in a position in which it cannot be observed.
In their study, Bertolotti and colleagues demonstrate that it is possible to form an image of an object hiding behind a scattering screen without the need to put a detector or a light beacon behind the screen. The authors placed a 50-micrometre-wide, two-dimensional fluorescent object at a distance of 6 mm behind a scattering screen, and shone laser light onto the screen. The light transmitted through the screen resulted in a random light pattern (speckle pattern), on the other side of the screen, that illuminated the fluorescent object. The researchers then measured the fluorescence that was generated by the object and transmitted back through the screen.
But how could they use these fluorescence measurements to form an image of the object, given that the speckle pattern illuminating the object was randomly generated by the scattering screen? The authors used fluorescence measurements of the object not only to form the image, but also as a beacon to probe the scattering medium8,9 so as to be able to undo its blurring effect on the image. First, they made multiple measurements by adjusting the angle of illumination slightly, thereby changing the unknown speckle pattern in a predictable way. Second, they repeated their experiment many times to obtain statistical averages of the properties of the scattering screen. These measurements supplied them with the information they needed to form the image.
This technique is capable of imaging objects some distance away (6 mm in the current study) from a thin scatterer (about 3–5 μm thick for ground glass). For example, it could be used to image two-dimensional fluorescent objects in blood or other liquids surrounded by a thin scattering layer. The approach will probably be extended to three-dimensional objects and possibly to non-fluorescent objects. However, major innovation would be required to expand the technique to permit imaging of objects behind or inside thick scattering media. For now, Bertolotti and colleagues' demonstration that it is possible to see clearly a fluorescent object behind a scattering screen, beyond the ballistic spatial limit, will almost certainly intensify the search for ways to use light to see through human tissue.
What does this story suggest for our golfer who wishes to hit the ball out of the woods and direct it towards the hole? She might have to hit many balls in various directions, the equivalent of adjusting the illumination angle in the authors' experiment. She might also have to engage many friends to stand around the golf course and shout back when a ball hits them, just as fluorescent molecules send light back when photons hit them. Even after hitting all these mulligans (second-chance shots in golf), she would not know how to strike the ball in the direction of the hole. However, if she hit enough of her friends during this unusual game, she would know where her friends were standing (the image of the object). But she would still not know how to hit a single shot through the trees and towards the hole — the analogue of ballistic passage through the scattering medium.
Bertolotti, J. et al. Nature 491, 232–234 (2012).
Wang, L., Ho, P., Liu, C., Zhang, G. & Alfano, R. Science 253, 769–771 (1991).
Alex, A. et al. J. Biomed. Opt. 15, 026025 (2010).
Helmchen, F. & Denk, W. Nature Methods 2, 932–940 (2005).
Leith, E. N. & Upatnieks, J. J. Opt. Soc. Am. 56, 523 (1966).
Kogelnik, H. & Pennington, K. S. J. Opt. Soc. Am. 58, 273–274 (1968).
Agarwal, G. S., Friberg, A. T. & Wolf, E. J. Opt. Soc. Am. 73, 529–537 (1983).
Vellekoop, I. M. & Mosk, A. P. Opt. Lett. 32, 2309–2311 (2007).
Hsieh, C. L., Pu, Y., Grange, R., Laporte, G. & Psaltis, D. Opt. Express 18, 20723–20731 (2010).
About this article
Scientific Reports (2018)
Optics Express (2015)
Nature Physics (2015)
Proceedings of the National Academy of Sciences (2015)