Misic, a general deep learning-based method for the high-throughput cell segmentation of complex bacterial communities.

Studies of bacterial communities, biofilms and microbiomes, are multiplying due to their impact on health and ecology. Live imaging of microbial communities requires new tools for the robust identification of bacterial cells in dense and often inter-species populations, sometimes over very large scales. Here, we developed MiSiC, a general deep-learning-based 2D segmentation method that automatically segments single bacteria in complex images of interacting bacterial communities with very little parameter adjustment, independent of the microscopy settings and imaging modality. Using a bacterial predator-prey interaction model, we demonstrate that MiSiC enables the analysis of interspecies interactions, resolving processes at subcellular scales and discriminating between species in millimeter size datasets. The simple implementation of MiSiC and the relatively low need in computing power make its use broadly accessible to fields interested in bacterial interactions and cell biology.

An all-optical technique enables instantaneous single-shot demodulation of images at high frequency.

High-frequency demodulation of wide area optical signals in a snapshot manner remains a technological challenge. If solved, it could open tremendous perspectives in 3D imaging, vibrometry, free-space communications, automated vision, or ballistic photon imaging in scattering media with numerous applications in smart autonomous vehicles and medical diagnosis. We present here a snapshot quadrature demodulation imaging technique, capable of estimating the amplitude and phase from a single acquisition, without synchronization of emitter and receiver, and with the added capability of continuous frequency tuning. This all-optical optimized setup comprises an electro-optic crystal acting as a fast sinusoidal optical transmission gate, and allows four quadrature image channels to be recorded simultaneously with any conventional camera. We report the design, experimental validation and examples of applications of such wide-field quadrature demodulating system that allowed snapshot demodulation of images with good spatial resolution and continuous frequency selectivity up to a few 100s of kilohertz.

Machine learning approach for rapid and accurate estimation of optical properties using spatial frequency domain imaging.

Fast estimation of optical properties from reflectance measurements at two spatial frequencies could pave way for real-time, wide-field and quantitative mapping of vital signs of tissues. We present a machine learning-based approach for estimating optical properties in the spatial frequency domain, where a random forest regression algorithm is trained over data obtained from Monte-Carlo photon transport simulations. The algorithm learns the nonlinear mapping between diffuse reflectance at two spatial frequencies, and the absorption and reduced scattering coefficient of the tissue under consideration. Using this method, absorption and reduced scattering properties could be obtained over a 1 megapixel image in 450 ms with errors as low as 0.556% in absorption and 0.126% in reduced scattering.

Theoretical optimal modulation frequencies for scattering parameter estimation and ballistic photon filtering in diffusing media.

The efficiency of using intensity modulated light for the estimation of scattering properties of a turbid medium and for ballistic photon discrimination is theoretically quantified in this article. Using the diffusion model for modulated photon transport and considering a noisy quadrature demodulation scheme, the minimum-variance bounds on estimation of parameters of interest are analytically derived and analyzed. The existence of a variance-minimizing optimal modulation frequency is shown and its evolution with the properties of the intervening medium is derived and studied. Furthermore, a metric is defined to quantify the efficiency of ballistic photon filtering which may be sought when imaging through turbid media. The analytical derivation of this metric shows that the minimum modulation frequency required to attain significant ballistic discrimination depends only on the reduced scattering coefficient of the medium in a linear fashion for a highly scattering medium.

Real-time imaging through strongly scattering media: seeing through turbid media, instantly.

Numerous everyday situations like navigation, medical imaging and rescue operations require viewing through optically inhomogeneous media. This is a challenging task as photons propagate predominantly diffusively (rather than ballistically) due to random multiple scattering off the inhomogenieties. Real-time imaging with ballistic light under continuous-wave illumination is even more challenging due to the extremely weak signal, necessitating voluminous data-processing. Here we report imaging through strongly scattering media in real-time and at rates several times the critical flicker frequency of the eye, so that motion is perceived as continuous. Two factors contributed to the speedup of more than three orders of magnitude over conventional techniques - the use of a simplified algorithm enabling processing of data on the fly, and the utilisation of task and data parallelization capabilities of typical desktop computers. The extreme simplicity of the technique, and its implementation with present day low-cost technology promises its utility in a variety of devices in maritime, aerospace, rail and road transport, in medical imaging and defence. It is of equal interest to the common man and adventure sportsperson like hikers, divers, mountaineers, who frequently encounter situations requiring realtime imaging through obscuring media. As a specific example, navigation under poor visibility is examined.

Long-range polarimetric imaging through fog.

We report an experimental implementation of long-range polarimetric imaging through fog over kilometric distance in real field atmospheric conditions. An incoherent polarized light source settled on a telecommunication tower is imaged at a distance of 1.3 km with a snapshot polarimetric camera including a birefringent Wollaston prism, allowing simultaneous acquisition of two images along orthogonal polarization directions. From a large number of acquisitions datasets and under various environmental conditions (clear sky/fog/haze, day/night), we compare the efficiency of using polarized light for source contrast increase with different signal representations (intensity, polarimetric difference, polarimetric contrast, etc.). With the limited-dynamics detector used, a maximum fourfold increase in contrast was demonstrated under bright background illumination using polarimetric difference image.

Optimal estimation in polarimetric imaging in the presence of correlated noise fluctuations.

We quantitatively analyze how a polarization-sensitive imager can overcome the precision of a standard intensity camera when estimating a parameter on a polarized source over an intense background. We show that the gain is maximized when the two polarimetric channels are perturbed with significantly correlated noise fluctuations. An optimal estimator is derived and compared to standard intensity and polarimetric estimators.

Enhancing the strength of an optical trap by truncation.

Optical traps (tweezers) are beginning to be used with increasing efficacy in diverse studies in the biological and biomedical sciences. We report here results of a systematic study aimed at enhancing the efficiency with which dielectric (transparent) materials can be optically trapped. Specifically, we investigate how truncation of the incident laser beam affects the strength of an optical trap in the presence of a circular aperture. Apertures of various sizes have been used by us to alter the beam radius, thereby changing the effective numerical aperture and intensity profile. We observe significant enhancement of the radial and axial trap stiffness when an aperture is used to truncate the beam compared to when no aperture was used, keeping incident laser power constant. Enhancement in trap stiffness persists even when the beam intensity profile is modulated. The possibility of applying truncation to multiple traps is explored; to this end a wire mesh is utilized to produce multiple trapping that also alters the effective numerical aperture. The use of a mesh leads to reduction in trap stiffness compared to the case when no wire mesh is used. Our findings lead to a simple-to-implement and inexpensive method of significantly enhancing optical trapping efficiency under a wide range of circumstances.