Accurate and practical feature extraction from noisy holograms

Characterization of Optical, Thermal, and Viscoelastic Properties of Pollenkitt in Angiosperm Pollen Using In-Line Digital Holographic Microscopy

Pollen grains are remarkable material composites, with various organelles in their fragile interior protected by a strong shell made of sporopollenin. The outermost layer of angiosperm pollen grains contains a lipid-rich substance called pollenkitt, which is a natural bioadhesive that helps preserve structural integrity when the pollen grain is exposed to external environmental stresses. In addition, its viscous nature enables it to adhere to various floral and insect surfaces, facilitating the pollination process. To examine the physicochemical properties of aqueous pollenkitt droplets, we used in-line digital holographic microscopy to capture light scattering from individual pollenkitt particles. Comparison of pollenkitt holograms to those modeled using the Lorenz-Mie theory enables investigations into the minute variations in the refractive index and size resulting from changes in local temperature and pollen aging.

cGAN-assisted imaging through stationary scattering media

Analyzing images taken through scattering media is challenging, owing to speckle decorrelations from perturbations in the media. For in-line imaging modalities, which are appealing because they are compact, require no moving parts, and are robust, negating the effects of such scattering becomes particularly challenging. Here we explore the use of conditional generative adversarial networks (cGANs) to mitigate the effects of the additional scatterers in in-line geometries, including digital holographic microscopy. Using light scattering simulations and experiments on objects of interest with and without additional scatterers, we find that cGANs can be quickly trained with minuscule datasets and can also efficiently learn the one-to-one statistical mapping between the cross-domain input-output image pairs. Importantly, the output images are faithful enough to enable quantitative feature extraction. We also show that with rapid training using only 20 image pairs, it is possible to negate this undesired scattering to accurately localize diffraction-limited impulses with high spatial accuracy, therefore transforming a shift variant system to a linear shift invariant (LSI) system.

Automated phase unwrapping in digital holography with deep learning

Digital holography can provide quantitative phase images related to the morphology and content of biological samples. After the numerical image reconstruction, the phase values are limited between -π and π; thus, discontinuity may occur due to the modulo 2π operation. We propose a new deep learning model that can automatically reconstruct unwrapped focused-phase images by combining digital holography and a Pix2Pix generative adversarial network (GAN) for image-to-image translation. Compared with numerical phase unwrapping methods, the proposed GAN model overcomes the difficulty of accurate phase unwrapping due to abrupt phase changes and can perform phase unwrapping at a twice faster rate. We show that the proposed model can generalize well to different types of cell images and has high performance compared to recent U-net models. The proposed method can be useful in observing the morphology and movement of biological cells in real-time applications.