Overcoming light scattering in bio-sample imaging

03 Jul 2014

Single-pixel optical system developed in Spain uses compressive sensing to see deeper inside tissues.

Optical imaging methods are rapidly becoming essential tools in biomedical science because they are noninvasive, fast, cost-efficient and pose no health risks because they do not use ionizing radiation.

These methods could become even more valuable if researchers could find a way for optical light to penetrate all the way through the body's tissues. With today’s technology, even passing through a fraction of an inch of skin is enough to scatter the light and scramble an image.

Now a team of researchers from Spain’s Jaume I University and the University of València has developed a single-pixel optical system based on compressive sensing that can overcome the fundamental limitations imposed by this scattering. The work was published in The Optical Society’s open-access journal Optics Express.

“In the diagnostic realm within the past few years, we have witnessed the way optical imaging has helped clinicians detect and evaluate suspicious lesions," said Jesús Lancis, the paper’s co-author and a researcher in the Photonics Research Group at UJI. "The elephant in the room, however, is the short penetration depth of light through tissue compared to ultrasound or X-ray techniques. Current knowledge is insufficient for early detection of small lesions located deeper than a millimeter beneath the surface of the mucosa. Our goal is to see more deeply inside tissue.”

Off-the-shelf system

The team use an off-the-shelf digital micromirror array from a commercial video projector to create a set of microstructured light patterns that are sequentially superimposed onto a sample. They then measure the transmitted energy with a photodetector that can sense the presence or absence of light, but has no spatial resolution. Then they apply a signal processing technique called compressive sensing, which is used to compress large data files as they are measured. This allows them to reconstruct the image.

One of the most surprising aspects of the team’s work is that they use essentially a single-pixel sensor to capture the images. While most people think that more pixels result in better image quality, there are some cases where this is not true, Lancis said. "In low-light imaging, for instance, it's better to integrate all available light into a single sensor. If the light is split into millions of pixels, each sensor receives a tiny fraction of light, creating noise and destroying the image."

“Something similar happens when you try to transmit images through scattering media,” he added. “When we use a conventional digital camera to get an image, we only see the familiar noisy pattern known as ‘speckle.’ In compressive imaging, since we aren’t using pixelated sensors, it should be less sensitive to light scrambling and enable transmission of images through scattering.”

Also notable, the technique could operate through dynamic scattering. “Most scattering media of interest, like biological tissues, are dynamic - such that the scatter centers continuously change their positions with time. This is ideal for some applications because monitoring the changes of the speckle can reveal information about the sample, but it’s a major nuisance for transmission or collection of images,” said Lancis. “Our technique, however, requires no calibration of the medium, and its fluctuations during the sensing stage don’t limit imaging ability.”

What’s ahead for the team? “Our next goal is to break the barriers of light penetration depth inside a scattering medium with the state-of-the-art megapixel-programmable spatial light modulators used in consumer electronics.”. To achieve this, they will need to demonstrate that the technique works even when the sample is embedded inside the tissue.

Conclusions

The Optics Express paper concludes, "We have demonstrated that computational techniques combined with single-pixel sensing enables image reconstruction behind arbitrary scattering media, in contrast to charge-coupled device cameras, where the pixelated structure of the sensor returns a noise-like speckle pattern. Our approach does not require a previous calibration of the disordered media and permits to retrieve images when we deal with dynamic scatterers.

"In contrast with techniques based on measuring the transmission matrix, our technique does not need to characterize the scattering medium, but operates on an intensity basis, thereby computing intensity distributions instead of complex fields. Moreover, the use of compressive sensing is limited to scenes that are sparse on the chosen basis.

"Our implementation is a first step to tackle the problem of imaging objects completely embedded in a scattering medium. In parallel, our disordered-assisted single-pixel configuration shows straightforward applications for image transmission through multimode fibers or to look around corners."

The principle of the new Spanish technique is presented in the following video:

About the Author

Matthew Peach is contributing editor to optics.org.