Imaging through scattering media

Using time-resolved data to predict an object’s form

Seeing “through” things may sound like something from science fiction, but Photon Force technologies have been successfully used to achieve just that.

Using time-resolved data captured with the Photon Force PF32 Time-Correlated Single-Photon Counting (TCSPC) camera, researchers have been able to predict the form of an object from behind a scattering medium or through tissue.

imaging through scattering media featured

How MIT Used The PF32 To Predict Form Behind A Barrier

Researchers at MIT have used the Photon Force PF32 to demonstrate an imaging technique that allows identification and classification of objects hidden behind scattering media.

The team used a machine learning algorithm to take the data from the PF32 camera and infer what’s concealed behind an opaque surface.

To the naked eye, the surface looks like a blur, however, having the very precise time-resolved information, knowing not only where the photons are, but when every photon is detected to 55ps accuracy, allowed the team to pull out more data from the image and reconstruct what was happening with the target that was hidden from view.

In the paper ‘Object classification through scattering media with deep learning on time-resolved measurement’, Satat et al. detail the data-driven approach and leveraging of convolutional neural networks (CNNs) that helped them create a model allowing robust imaging through scattering conditions that is not sensitive to calibration.

In this figure (a) shows the training of the CNN and (b) shows the physical set up of the pulsed laser and the PF32 SPAD array camera.

By analysing (a) the time-resolved data from the PF32 of an unknown object form, the researchers were able to (b) predict the classification of the figure’s pose with a high degree of accuracy

How Proteus used SPAD cameras to aid optical endomicroscopy

The aim of Proteus is to develop a revolutionary technology to provide quick, bedside diagnosis and management of lung diseases in the clinical environment. An Interdisciplinary Research Collaboration (IRC), Proteus is funded by EPSRC and is a consortium between the Universities of Edinburgh, Bath and Heriot-Watt.

In one piece of work, Proteus team members have used SPAD cameras to locate optical endomicroscopy fibres inside organic tissue. In the paper ‘Ballistic and snake photon imaging for locating optical enomicroscopy fibres’ by Tanner et al, the team demonstrate determination of the location of the distal-end of a fibre-optic device deep in tissue through the imaging of ballistic and snake photons, using a time-resolved single-photon detector array.

Using the set-up shown in (a) to (c), the team noted that the pulsed light that exits the fibre is mostly scattered by the tissue, however, some photons pass through without scattering and are known as ballistic photons.

In diagram (d) at t=X, the observation is of the ballistic photons arriving at the camera. At times beyond this the camera receives light that has been scattered within the tissue. Using the arrival time of the ballistic photons correlated with the time at which the laser pulse was emitted, the team could ascertain where the fibre is.