Smartphones Yield 3D Images on the Cheap

A scientist in Japan has devised a new way of using a smartphone to capture 3D images in the form of holograms (Appl. Opt., doi: doi.org/10.1364/AO.532972). He says that his scheme, which relies on compressing the data within a digital hologram, is compact and cheap and might prove useful for medical diagnosis―particularly in remote areas.

3D reconstructions

The idea underlying the work is to use a digital form of holography to obtain both amplitude and phase information from a microscopic sample in a single shot. Like other forms of holography, this involves bouncing light off the specimen and interfering the scattered beam with a reference beam. Holograms are then reconstructed by using wave propagation calculations, allowing the object to be viewed in three dimensions.

Researchers have already shown how this principle could be exploited for bio-imaging, to study cell division and other cell dynamics as well as diagnose sickle cell disease. But most demonstrations to date have been one-offs, involving bulky and expensive equipment.

Some groups have instead used smartphones as part of the process of collecting and reconstructing images. However, these systems all have their limitations―they either require a separate device for reconstructing the holograms or forego real-time image generation by doing everything on the phone.

There’s an app for that

In the latest work, Yuki Nagahama at the Tokyo University of Agriculture and Technology has found it is possible to reconstruct holograms on a smartphone in close to real time using an Android-based application that he designed. The system relies on a separate optical device to record the holograms, but this device can be made using a compact 3D printer and contains nothing more than a semiconductor laser and standard USB camera—neither bulky nor expensive. (Using the smartphone itself to do the recording would mean removing the lens from the phone’s camera, which would prevent it from taking pictures.)

Key to the new scheme is a less computation-intensive form of digital holography better suited to smartphones, which tend to have limited processing and memory capabilities.

Key to the new scheme is a less computation-intensive form of digital holography better suited to smartphones, which tend to have limited processing and memory capabilities. This “band-limited double-step Fresnel diffraction” involves creating a virtual plane between the hologram and image planes, and then placing that imaginary surface at whatever distance is required to yield a given sampling rate at the image plane (placing this surface nearer the object lowers the required rate but also limits the area that can be observed).

Reproducing a pine leaf

To put his smartphone-based system through its paces, Nagahama used it to image a leaf from a pine tree. He showed that it could accurately reproduce both the object's amplitude and phase information, and he found that while holography using convolution-based diffraction yielded about 0.7 frames per second, the double-step method instead reached about 1.9 frames per second.

What's more, he also showed it was possible to zoom in on and out of the reconstructed image by pinching the phone's touch screen. He enabled this feature by using what are known as scalable diffraction calculations.

Nagahama says that his next step is to use deep-learning techniques to remove the unwanted secondary images that are sometimes produced when reconstructing digital holograms. Eventually, he hopes to capture holograms of moving as well as static objects by employing multiple processors to carry out calculations in parallel―perhaps using the graphics processing unit built into a smartphone’s system on a chip. As for how much the system might cost, he says that the most expensive material item is the USB camera, which, he points out, costs just a few tens of dollars.