Search

We present a novel deep learning approach to reconstruct confocal microscopy stacks from single light field images. To perform the reconstruction, we introduce LFMNet, a fully convolutional neural network architecture inspired by the U-Net design. A property of LFMNet is its ability to reconstruct with near confocal accuracy a $112 \times 112 \times 57.6 \,\mu {\rm m}^3$ volume ( $1287\times 1287\times 64$ voxels) in $\text{50}\,ms$ given a single light field image of $1287\times 1287$ pixels, thus dramatically reducing 720-fold the capture time compared to confocal scanning of assays at the same volumetric resolution and 64-fold the required storage. To prove its applicability in life sciences, our approach is evaluated both quantitatively and qualitatively on mouse brain slices with fluorescently labeled blood vessels. Because of the drastic reduction in image acquisition time, our setup and method are suitable for imaging highly dynamic and light-sensitive events, and in closed-loop systems, where latency due to the reconstruction time is crucial. We provide empirical analysis of the optical design, of the network architecture and of our training procedure to reconstruct confocal-like volumes for a given target depth range. To train our network, we built a data set of 362 light field images of the mouse brain sample and the corresponding aligned set of 3D confocal scans, which we use as ground truth. The data set will be made available for research purposes.

Collagen is the main structural element of connective tissues, and its favorable properties make it an ideal biomaterial for regenerative medicine. In dental medicine, collagen barrier membranes fabricated from naturally occurring tissues are used for guided bone regeneration. Since the morphological characteristics of collagen membranes play a crucial role in their mechanical properties and affect the cellular behavior at the defect site, in-depth knowledge of the structure is key. As a base for the development of novel collagen membranes, an extensive morphological analysis of four porcine membranes, including centrum tendineum, pericardium, plica venae cavae and small intestinal submucosa, was performed. Native membranes were analyzed in terms of their thickness. Second harmonic generation and two-photon excitation microscopy of the native membranes showed the 3D architecture of the collagen and elastic fibers, as well as a volumetric index of these two membrane components. The surface morphology, fiber arrangement, collagen fibril diameter and D-periodicity of decellularized membranes were investigated by scanning electron microscopy. All the membrane types showed significant differences in thickness. In general, undulating collagen fibers were arranged in stacked layers, which were parallel to the membrane surface. Multiphoton microscopy revealed a conspicuous superficial elastic fiber network, while the elastin content in deeper layers varied. The elastin/collagen volumetric index was very similar in the investigated membranes and indicated that the collagen content was clearly higher than the elastin content. The surface of both the pericardium and plica venae cavae and the cranial surface of the centrum tendineum revealed a smooth, tightly arranged and crumpled morphology. On the caudal face of the centrum tendineum, a compact collagen arrangement was interrupted by clusters of circular discontinuities. In contrast, both surfaces of the small intestinal submucosa were fibrous, fuzzy and irregular. All the membranes consisted of largely uniform fibrils displaying the characteristic D-banding. This study reveals similarities and relevant differences among the investigated porcine membranes, suggesting that each membrane represents a unique biomaterial suitable for specific applications.