Root CT Segmentation Using Incremental Learning Methodology on Improved Multiple Resolution Images

The real-time issue of reliability segmenting root structure while using X-Ray Computed Tomography (CT) images is addressed in this work. A deep learning approach is proposed using a novel framework, involving decoders and encoders. The encoders-decoders framework is useful to improve multiple resolution by means of upsampling and downsampling images. The methodology of the work is enhanced by incorporating network branches with individual tasks using low-resolution context information and high-resolution segmentation. In large volumetric images, it is possible to resolve small root details by implementing a memory efficient system, resulting in the formation of a complete network. The proposed work, recent image analysis tool developed for root CT segmented is compared with several other previously existing methodology and it is found that this methodology is more efficient. Quantitatively and qualitatively, it is found that a multiresolution approach provides high accuracy in a shallower network with a large receptive field or deep network in a small receptive field. An incremental learning approach is also embedded to enhance the performance of the system. Moreover, it is also capable of detecting fine and large root materials in the entire volume. The proposed work is fully automated and doesn’t require user interaction.

4D reconstruction of developmental trajectories using spherical harmonics

Although the full embryonic development of species such as Drosophila and zebrafish can be 3D imaged in real time, this is not true for mammalian organs, as normal organogenesis cannot be recapitulated in vitro. Currently available 3D data is therefore ex vivo images which provide only a snap shot of development at discrete moments in time. Here we propose a computer based approach to recreate the continuous evolution in time and space of developmental stages from 3D volumetric images. Our method uses the mathematical approach of spherical harmonics to re-map discrete shape data into a space in which facilitates a smooth interpolation over time. We tested our approach on mouse limb buds (from E10 to E12.5) and embryonic hearts (from 10 to 29 somites). A key advantage of the method is that the resulting 4D trajectory takes advantage of all the available data (i.e. it is not dominated by the choice of a few "ideal" images), while also being able to interpolate well through time intervals for which there is little or no data. This method not only provides a quantitative basis for validating predictive models, but it also increases our understanding of morphogenetic processes. We believe this is the first data-driven quantitative 4D description of limb morphogenesis.

Growth Kinetics of Single Polymer Particles in Solution via Active-Feedback 3D Tracking

The ability to directly observe chemical reactions at the single-molecule and single-particle level has enabled the discovery of behaviors otherwise obscured by the ensemble averaging in bulk measurements. However powerful, a common restriction of these studies to date has been the absolute requirement to surface tether or otherwise immobilize the chemical reagent/reaction of interest. This constraint arose from a fundamental limitation of conventional microscopy techniques, which could not track molecules or particles rapidly diffusing in three dimensions, as occurs in solution. However, much chemistry occurs in the solution phase, leaving single-particle/-molecule analysis of this critical area of science beyond the scope of available technology. Here we report the first solution-phase studies and measurements of any chemical reaction at single-particle/-molecule level in freely diffusing solution. During chemical reaction, freely diffusing polymer particles (D ~ 10-12 m2/s) yielded single-particle 3D trajectories and real-time volumetric images that were analyzed to extract the growth rates of individual particles. These volumetric images show that the average growth rate is a poor representation of the true underlying variability in polymer-particle growth behavior. These data revealed statistically significant populations of faster- and slower-growing particles at different depths in the sample, showing emergent heterogeneity while particles are still in the solution phase. These results go against the prevailing premise that chemical processes freely diffusing in solution will exhibit uniform kinetics. These new understandings of mechanisms behind polymer growth variations bring about an exciting opportunity to control particle-size and plausibly molecular weight polydispersity by the rational design of conditions to dictate spatial growth gradients. We anticipate that these studies will launch a new field of solution-phase, nonensemble-averaged measurements of chemical reactions.

Automated segmentation of microtomography imaging of Egyptian mummies

Propagation Phase Contrast Synchrotron Microtomography (PPC-SRμCT) is the gold standard for non-invasive and non-destructive access to internal structures of archaeological remains. In this analysis, the virtual specimen needs to be segmented to separate different parts or materials, a process that normally requires considerable human effort. In the Automated SEgmentation of Microtomography Imaging (ASEMI) project, we developed a tool to automatically segment these volumetric images, using manually segmented samples to tune and train a machine learning model. For a set of four specimens of ancient Egyptian animal mummies we achieve an overall accuracy of 94–98% when compared with manually segmented slices, approaching the results of off-the-shelf commercial software using deep learning (97–99%) at much lower complexity. A qualitative analysis of the segmented output shows that our results are close in terms of usability to those from deep learning, justifying the use of these techniques.

Post hoc Correction of Chromatic Aberrations in Large-Scale Volumetric Images in Confocal Microscopy

Over the last decade, tissue-clearing techniques have expanded the scale of volumetric fluorescence imaging of the brain, allowing for the comprehensive analysis of neuronal circuits at a millimeter scale. Multicolor imaging is particularly powerful for circuit tracing with fluorescence microscopy. However, multicolor imaging of large samples often suffers from chromatic aberration, where different excitation wavelengths of light have different focal points. In this study, we evaluated chromatic aberrations for representative objective lenses and a clearing agent with confocal microscopy and found that axial aberration is particularly problematic. Moreover, the axial chromatic aberrations were often depth-dependent. Therefore, we developed a program that is able to align depths for different fluorescence channels based on reference samples with fluorescent beads or data from guide stars within biological samples. We showed that this correction program can successfully correct chromatic aberrations found in confocal images of multicolor-labeled brain tissues. Our simple post hoc correction strategy is useful to obtain large-scale multicolor images of cleared tissues with minimal chromatic aberrations.

Deep learning-based segmentation of lithium-ion battery microstructures enhanced by artificially generated electrodes

Accurate 3D representations of lithium-ion battery electrodes, in which the active particles, binder and pore phases are distinguished and labeled, can assist in understanding and ultimately improving battery performance. Here, we demonstrate a methodology for using deep-learning tools to achieve reliable segmentations of volumetric images of electrodes on which standard segmentation approaches fail due to insufficient contrast. We implement the 3D U-Net architecture for segmentation, and, to overcome the limitations of training data obtained experimentally through imaging, we show how synthetic learning data, consisting of realistic artificial electrode structures and their tomographic reconstructions, can be generated and used to enhance network performance. We apply our method to segment x-ray tomographic microscopy images of graphite-silicon composite electrodes and show it is accurate across standard metrics. We then apply it to obtain a statistically meaningful analysis of the microstructural evolution of the carbon-black and binder domain during battery operation.

In the Era of Systemic Therapy for Hepatocellular Carcinoma Is Transarterial Chemoembolization Still a Card to Play?

Conventional transarterial embolization (cTACE) has been proven to be effective for intermediate stage hepatocellular carcinoma (HCC), with a recent systematic review showing an overall survival (OS) of 19.4 months. Nevertheless, due to the rapid development of the systemic therapeutic landscape, the place of TACE is becoming questionable. Is there still a niche for TACE in the era of immunotherapy and combination treatments such as atezolizumab–bevacizumab, which has shown an OS of 19.2 months with excellent tolerance? The development of drug-eluting microspheres (DEMs) has led to the standardization of the technique, and along with adequate selection, it showed an OS of 48 months in a retrospective study. In order to increase treatment selectivity, new catheters have also been added to the TACE arsenal as well as the use of cone-beam CT (CBCT), which provides three-dimensional volumetric images and guidance during procedures. Moreover, the TACE indications have also widened. It may serve as a “bridging therapy” for liver transplantation candidates while they are on the waiting list, and it represents a valuable downstaging tool to transplantation criteria. The aim of this review is to explore the current data on the advancements of TACE and its future place amongst the growing panel of treatments.

Hands-On Deformation of Volumetric Anatomical Imageson a Touchscreen

In this work, we propose a novel metaphor to interact with volumetric anatomical images, e.g., magnetic resonance imaging or computed tomography scans. Beyond simple visual inspection, we empower users to reach the visible anatomical elements directly with their hands, and then move and deform them through natural gestures, while respecting the mechanical behavior of the underlying anatomy. This interaction metaphor relies on novel technical methods that address three major challenges: selection of anatomical elements in volumetric images, mapping of 2D manipulation gestures to 3D transformations, and real-time deformation of the volumetric images. All components of the interaction metaphor have been designed to capture the user’s intent in an intuitive manner, solving the mapping from the 2D touchscreen to the visible elements of the 3D volume. As a result, users have the ability to interact with medical volume images much like they would do with physical anatomy, directly with their hands.