Alternation of inverse problem approach and deep learning for lens-free microscopy image reconstruction

A lens-free microscope is a simple imaging device performing in-line holographic measurements. In the absence of focusing optics, a reconstruction algorithm is used to retrieve the sample image by solving the inverse problem. This is usually performed by optimization algorithms relying on gradient computation. However the presence of local minima leads to unsatisfactory convergence when phase wrapping errors occur. This is particularly the case in large optical thickness samples, for example cells in suspension and cells undergoing mitosis. To date, the occurrence of phase wrapping errors in the holographic reconstruction limits the application of lens-free microscopy in live cell imaging. To overcome this issue, we propose a novel approach in which the reconstruction alternates between two approaches, an inverse problem optimization and deep learning. The computation starts with a first reconstruction guess of the cell sample image. The result is then fed into a neural network, which is trained to correct phase wrapping errors. The neural network prediction is next used as the initialization of a second and last reconstruction step, which corrects to a certain extent the neural network prediction errors. We demonstrate the applicability of this approach in solving the phase wrapping problem occurring with cells in suspension at large densities. This is a challenging sample that typically cannot be reconstructed without phase wrapping errors, when using inverse problem optimization alone.

Multispectral total-variation reconstruction applied to lens-free microscopy

Lens-free microscopy multispectral acquisitions are processed with an inverse problem approach: a multispectral total variation criterion is defined and minimized with the conjugate gradients method. Reconstruction results show that the method is efficient to recover the phase image of densely packed cells.

Dynamics of cell and tissue growth acquired by means of extended field of view lensfree microscopy

In this paper, we discuss a new methodology based on lensfree imaging to perform wound healing assay with unprecedented statistics. Our video lensfree microscopy setup is a simple device featuring only a CMOS sensor and a semi coherent illumination system. Yet it is a powerful mean for the real-time monitoring of cultivated cells. It presents several key advantages, e.g. integration into standard incubator, compatibility with standard cell culture protocol, simplicity and ease of use. It can perform the follow-up in a large field of view (25 mm(2)) of several crucial parameters during the culture of cells i.e. their motility, their proliferation rate or their death. Consequently the setup can gather large statistics both in space and time. Here we uses this facility in the context of wound healing assay to perform label-free measurements of the velocities of the fronts of proliferation of the cell layer as a function of time by means of particle image velocimetry (PIV) processing. However, for such tissue growth experiments, the field of view of 25 mm(2) remains not sufficient and results can be biased depending on the position of the device with respect to the recipient of the cell culture. Hence, to conduct exhaustive wound healing assays, we propose to enlarge the field of view up to 10 cm(2) through a raster scan, by moving the source/sensor with respect to the Petri dish. We have performed acquisitions of wound healing assay (keratinocytes HaCaT) both in real-time (25 mm(2)) and in final point (10 cm(2)) to assess the combination of velocimetry measurements and final point wide field imaging. In the future, we aim at combining directly our extended field of view acquisitions (>10 cm(2)) with real time ability inside the incubator.