DeepSSM: A blueprint for image-to-shape deep learning models.

Statistical shape modeling (SSM) characterizes anatomical variations in a population of shapes generated from medical images. Statistical analysis of shapes requires consistent shape representation across samples in shape cohort. Establishing this representation entails a processing pipeline that includes anatomy segmentation, image re-sampling, shape-based registration, and non-linear, iterative optimization. These shape representations are then used to extract low-dimensional shape descriptors that are anatomically relevant to facilitate subsequent statistical analyses in different applications. However, the current process of obtaining these shape descriptors from imaging data relies on human and computational resources, requiring domain expertise for segmenting anatomies of interest. Moreover, this same taxing pipeline needs to be repeated to infer shape descriptors for new image data using a pre-trained/existing shape model. Here, we propose DeepSSM, a deep learning-based framework for learning the functional mapping from images to low-dimensional shape descriptors and their associated shape representations, thereby inferring statistical representation of anatomy directly from 3D images. Once trained using an existing shape model, DeepSSM circumvents the heavy and manual pre-processing and segmentation required by classical models and significantly improves the computational time, making it a viable solution for fully end-to-end shape modeling applications. In addition, we introduce a model-based data-augmentation strategy to address data scarcity, a typical scenario in shape modeling applications. Finally, this paper presents and analyzes two different architectural variants of DeepSSM with different loss functions using three medical datasets and their downstream clinical application. Experiments showcase that DeepSSM performs comparably or better to the state-of-the-art SSM both quantitatively and on application-driven downstream tasks. Therefore, DeepSSM aims to provide a comprehensive blueprint for deep learning-based image-to-shape models.

High-entropy oxychloride increasing the stability of Li-sulfur batteries.

A novel lithiated high-entropy oxychloride Li0.5(Zn0.25Mg0.25Co0.25Cu0.25)0.5Fe2O3.5Cl0.5 (LiHEOFeCl) with spinel structure belonging to the cubic Fd3̄m space group is synthesized by a mechanochemical-thermal route. Cyclic voltammetry measurement of the pristine LiHEOFeCl sample confirms its excellent electrochemical stability and the initial charge capacity of 648 mA h g-1. The reduction of LiHEOFeCl starts at ca. 1.5 V vs. Li+/Li, which is outside the electrochemical window of the Li-S batteries (1.7/2.9 V). The addition of the LiHEOFeCl material to the composite of carbon with sulfur results in improved long-term electrochemical cycling stability and increased charge capacity of this cathode material in Li-S batteries. The carbon/LiHEOFeCl/sulfur cathode provides a charge capacity of 530 mA h g-1 after 100 galvanostatic cycles, which represents ca. 33% increase as compared to the charge capacity of the blank carbon/sulfur composite cathode after 100 cycles. This considerable effect of the LiHEOFeCl material is assigned to its excellent structural and electrochemical stability within the potential window of 1.7 V/2.9 V vs. Li+/Li. In this potential region, our LiHEOFeCl has no inherent electrochemical activity. Hence, it acts solely as an electrocatalyst accelerating the redox reactions of polysulfides. This can be beneficial for the performance of Li-S batteries, as evidenced by reference experiments with TiO2 (P90).

Quantifying the Severity of Metopic Craniosynostosis Using Unsupervised Machine Learning.

BACKGROUND: Quantifying the severity of head shape deformity and establishing a threshold for operative intervention remains challenging in patients with metopic craniosynostosis (MCS). This study combines three-dimensional skull shape analysis with an unsupervised machine-learning algorithm to generate a quantitative shape severity score (cranial morphology deviation) and provide an operative threshold score. METHODS: Head computed tomography scans from subjects with MCS and normal controls (5 to 15 months of age) were used for objective three-dimensional shape analysis using ShapeWorks software and in a survey for craniofacial surgeons to rate head-shape deformity and report whether they would offer surgical correction based on head shape alone. An unsupervised machine-learning algorithm was developed to quantify the degree of shape abnormality of MCS skulls compared to controls. RESULTS: One hundred twenty-four computed tomography scans were used to develop the model; 50 (24% MCS, 76% controls) were rated by 36 craniofacial surgeons, with an average of 20.8 ratings per skull. The interrater reliability was high (intraclass correlation coefficient, 0.988). The algorithm performed accurately and correlates closely with the surgeons assigned severity ratings (Spearman correlation coefficient, r = 0.817). The median cranial morphology deviation for affected skulls was 155.0 (interquartile range, 136.4 to 194.6; maximum, 231.3). Skulls with ratings of 150.2 or higher were very likely to be offered surgery by the experts in this study. CONCLUSIONS: This study describes a novel metric to quantify the head shape deformity associated with MCS and contextualizes the results using clinical assessments of head shapes by craniofacial experts. This metric may be useful in supporting clinical decision making around operative intervention and in describing outcomes and comparing patient population across centers.

Oxidative dehydrogenation of cyclohexene on atomically precise subnanometer Cu4-nPdn (0 ≤ n ≤ 4) tetramer clusters: the effect of cluster composition and support on performance.

The pronounced effects of the composition of four-atom monometallic Cu and Pd and bimetallic CuPd clusters and the support on the catalytic activity and selectivity in the oxidative dehydrogenation of cyclohexene are reported. The ultra-nanocrystalline diamond supported clusters are highly active and dominantly produce benzene; some of the mixed clusters also produce cyclohexadiene, which are all clusters with a much suppressed combustion channel. The also highly active TiO2-supported tetramers solely produce benzene, without any combustion to CO2. The selectivity of the zirconia-supported mixed CuPd clusters and the monometallic Cu cluster is entirely different; though they are less active in comparison to clusters with other supports, these clusters produce significant fractions of cyclohexadiene, with their selectivity towards cyclohexadiene gradually increasing with the increasing number of copper atoms in the cluster, reaching about 50% for Cu3Pd1. The zirconia-supported copper tetramer stands out from among all the other tetramers in this reaction, with a selectivity towards cyclohexadiene of 70%, which far exceeds those of all the other cluster-support combinations. The findings from this study indicate a positive effect of copper on the stability of the mixed tetramers and potential new ways of fine-tuning catalyst performance by controlling the composition of the active site and via cluster-support interactions in complex oxidative reactions under the suppression of the undesired combustion of the feed.

Atom by atom built subnanometer copper cluster catalyst for the highly selective oxidative dehydrogenation of cyclohexene.

The effect of particle size and support on the catalytic performance of supported subnanometer copper clusters was investigated in the oxidative dehydrogenation of cyclohexene. From among the investigated seven size-selected subnanometer copper particles between a single atom and clusters containing 2-7 atoms, the highest activity was observed for the titania-supported copper tetramer with 100% selectivity toward benzene production and being about an order of magnitude more active than not only all the other investigated cluster sizes on the same support but also the same tetramer on the other supports, Al2O3, SiO2, and SnO2. In addition to the profound effect of cluster size on activity and with Cu4 outstanding from the studied series, Cu4 clusters supported on SiO2 provide an example of tuning selectivity through support effects when this particular catalyst also produces cyclohexadiene with about 30% selectivity. Titania-supported Cu5 and Cu7 clusters supported on TiO2 produce a high fraction of cyclohexadiene in contrast to their neighbors, while Cu4 and Cu6 solely produce benzene without any combustion, thus representing odd-even oscillation of selectivity with the number of atoms in the cluster.

Interactive Deep Colorization and its Application for Image Compression.

Recent methods based on deep learning have shown promise in converting grayscale images to colored ones. However, most of them only allow limited user inputs (no inputs, only global inputs, or only local inputs), to control the output colorful images. The possible difficulty lies in how to differentiate the influences of different inputs. To solve this problem, we propose a two-stage deep colorization method allowing users to control the results by flexibly setting global inputs and local inputs. The key steps include enabling color themes as global inputs by extracting K mean colors and generating K-color maps to define a global theme loss, and designing a loss function to differentiate the influences of different inputs without causing artifacts. We also propose a color theme recommendation method to help users choose color themes. Based on the colorization model, we further propose an image compression scheme, which supports variable compression ratios in a single network. Experiments on colorization show that our method can flexibly control the colorized results with only a few inputs and generate state-of-the-art results. Experiments on compression show that our method achieves much higher image quality at the same compression ratio when compared to the state-of-the-art methods.

Soft Articulated Characters in Projective Dynamics.

We propose a fast and robust solver to simulate continuum-based deformable models with constraints, in particular, rigid-body and joint constraints useful for soft articulated characters. Our method embeds the degrees of freedom of both articulated rigid bodies and deformable bodies in one unified constrained optimization problem, thus coupling the deformable and rigid bodies. Inspired by Projective Dynamics which is a fast numerical solver to simulate deformable objects, we also propose a novel local/global solver that takes full advantage of the pre-factorized system matrices to accelerate the solve of our constrained optimization problem. Therefore, our method can efficiently simulate character models, with rigid-body parts (bones) being correctly coupled with deformable parts (flesh). Our method is stable because backward Euler time integration is applied to both rigid and deformable degrees of freedom. Our unified optimization problem is rigorously derived from constrained Newtonian mechanics. When simulating only articulated rigid bodies as a special case, our method converges to the state-of-the-art rigid body simulators.

Leveraging unsupervised image registration for discovery of landmark shape descriptor.

In current biological and medical research, statistical shape modeling (SSM) provides an essential framework for the characterization of anatomy/morphology. Such analysis is often driven by the identification of a relatively small number of geometrically consistent features found across the samples of a population. These features can subsequently provide information about the population shape variation. Dense correspondence models can provide ease of computation and yield an interpretable low-dimensional shape descriptor when followed by dimensionality reduction. However, automatic methods for obtaining such correspondences usually require image segmentation followed by significant preprocessing, which is taxing in terms of both computation as well as human resources. In many cases, the segmentation and subsequent processing require manual guidance and anatomy specific domain expertise. This paper proposes a self-supervised deep learning approach for discovering landmarks from images that can directly be used as a shape descriptor for subsequent analysis. We use landmark-driven image registration as the primary task to force the neural network to discover landmarks that register the images well. We also propose a regularization term that allows for robust optimization of the neural network and ensures that the landmarks uniformly span the image domain. The proposed method circumvents segmentation and preprocessing and directly produces a usable shape descriptor using just 2D or 3D images. In addition, we also propose two variants on the training loss function that allows for prior shape information to be integrated into the model. We apply this framework on several 2D and 3D datasets to obtain their shape descriptors. We analyze these shape descriptors in their efficacy of capturing shape information by performing different shape-driven applications depending on the data ranging from shape clustering to severity prediction to outcome diagnosis.

In Situ Raman Microdroplet Spectroelectrochemical Investigation of CuSCN Electrodeposited on Different Substrates.

Systematic in situ Raman microdroplet spectroelectrochemical (Raman-µSEC) characterization of copper (I) thiocyanate (CuSCN) prepared using electrodeposition from aqueous solution on various substrates (carbon-based, F-doped SnO2) is presented. CuSCN is a promising solid p-type inorganic semiconductor used in perovskite solar cells as a hole-transporting material. SEM characterization reveals that the CuSCN layers are homogenous with a thickness of ca. 550 nm. Raman spectra of dry CuSCN layers show that the SCN- ion is predominantly bonded in the thiocyanate resonant form to copper through its S-end (Cu-S-C≡N). The double-layer capacitance of the CuSCN layers ranges from 0.3 mF/cm2 on the boron-doped diamond to 0.8 mF/cm2 on a glass-like carbon. In situ Raman-µSEC shows that, independently of the substrate type, all Raman vibrations from CuSCN and the substrate completely vanish in the potential range from 0 to -0.3 V vs. Ag/AgCl, caused by the formation of a passivation layer. At positive potentials (+0.5 V vs. Ag/AgCl), the bands corresponding to the CuSCN vibrations change their intensities compared to those in the as-prepared, dry layers. The changes concern mainly the Cu-SCN form, showing the dependence of the related vibrations on the substrate type and thus on the local environment modifying the delocalization on the Cu-S bond.

Nanocrystalline TiO2/Carbon/Sulfur Composite Cathodes for Lithium-Sulfur Battery.

This paper evaluates the influence of the morphology, surface area, and surface modification of carbonaceous additives on the performance of the corresponding cathode in a lithium-sulfur battery. The structure of sulfur composite cathodes with mesoporous carbon, activated carbon, and electrochemical carbon is studied by X-ray diffraction, nitrogen adsorption measurements, and Raman spectroscopy. The sulfur cathode containing electrochemical carbon with the specific surface area of 1606.6 m2 g-1 exhibits the best electrochemical performance and provides a charge capacity of almost 650 mAh g-1 in cyclic voltammetry at a 0.1 mV s-1 scan rate and up to 1300 mAh g-1 in galvanostatic chronopotentiometry at a 0.1 C rate. This excellent electrochemical behavior is ascribed to the high dispersity of electrochemical carbon, enabling a perfect encapsulation of sulfur. The surface modification of carbonaceous additives by TiO2 has a positive effect on the electrochemical performance of sulfur composites with mesoporous and activated carbons, but it causes a loss of dispersity and a consequent decrease of the charge capacity of the sulfur composite with electrochemical carbon. The composite of sulfur with TiO2-modified activated carbon exhibited the charge capacity of 393 mAh g-1 in cyclic voltammetry and up to 493 mAh g-1 in galvanostatic chronopotentiometry. The presence of an additional Sigracell carbon felt interlayer further improves the electrochemical performance of cells with activated carbon, electrochemical carbon, and nanocrystalline TiO2-modified activated carbon. This positive effect is most pronounced in the case of activated carbon modified by nanocrystalline TiO2. However, it is not boosted by additional coverage by TiO2 or SnO2, which is probably due to the blocking of pores.

Atomic layer deposited films of Al2O3 on fluorine-doped tin oxide electrodes: stability and barrier properties.

Al2O3 layers were deposited onto electrodes by atomic layer deposition. Solubility and electron-transport blocking were tested. Films deposited onto fluorine-doped tin oxide (FTO, F:SnO2/glass) substrates blocked electron transfer to redox couples (ferricyanide/ferrocyanide) in aqueous media. However, these films were rapidly dissolved in 1 M NaOH (≈100 nm/h). The dissolution was slower in 1 M H2SO4 (1 nm/h) but after 24 h the blocking behaviour was entirely lost. The optimal stability was reached at pH 7.2 where no changes were found up to 24 h and even after 168 h of exposure the changes in the blocking behaviour were still minimal. This behaviour was also observed for protection against direct reduction of FTO.

Selected Electrochemical Properties of 4,4'-((1E,1'E)-((1,2,4-Thiadiazole-3,5-diyl)bis(azaneylylidene))bis(methaneylylidene))bis(N,N-di-p-tolylaniline) towards Perovskite Solar Cells with 14.4% Efficiency.

Planar perovskite solar cells were fabricated on F-doped SnO2 (FTO) coated glass substrates, with 4,4'-((1E,1'E)-((1,2,4-thiadiazole-3,5-diyl)bis(azaneylylidene))bis(methaneylylidene))bis(N,N-di-p-tolylaniline) (bTAThDaz) as hole transport material. This imine was synthesized in one step reaction, starting from commercially available and relatively inexpensive reagents. Electrochemical, optical, electrical, thermal and structural studies including thermal images and current-voltage measurements of the full solar cell devices characterize the imine in details. HOMO-LUMO of bTAThDaz were investigated by cyclic voltammetry (CV) and energy-resolved electrochemical impedance spectroscopy (ER-EIS) and were found at -5.19 eV and -2.52 eV (CV) and at -5.5 eV and -2.3 eV (ER-EIS). The imine exhibited 5% weight loss at 156 °C. The electrical behavior and photovoltaic performance of the perovskite solar cell was examined for FTO/TiO2/perovskite/bTAThDaz/Ag device architecture. Constructed devices exhibited good time and air stability together with quite small effect of hysteresis. The observed solar conversion efficiency was 14.4%.

Quantifying the Severity of Metopic Craniosynostosis: A Pilot Study Application of Machine Learning in Craniofacial Surgery.

The standard for diagnosing metopic craniosynostosis (CS) utilizes computed tomography (CT) imaging and physical exam, but there is no standardized method for determining disease severity. Previous studies using interfrontal angles have evaluated differences in specific skull landmarks; however, these measurements are difficult to readily ascertain in clinical practice and fail to assess the complete skull contour. This pilot project employs machine learning algorithms to combine statistical shape information with expert ratings to generate a novel objective method of measuring the severity of metopic CS.Expert ratings of normal and metopic skull CT images were collected. Skull-shape analysis was conducted using ShapeWorks software. Machine-learning was used to combine the expert ratings with our shape analysis model to predict the severity of metopic CS using CT images. Our model was then compared to the gold standard using interfrontal angles.Seventeen metopic skull CT images of patients 5 to 15 months old were assigned a severity by 18 craniofacial surgeons, and 65 nonaffected controls were included with a 0 severity. Our model accurately correlated the level of skull deformity with severity (P < 0.10) and predicted the severity of metopic CS more often than models using interfrontal angles (χ = 5.46, P = 0.019).This is the first study that combines shape information with expert ratings to generate an objective measure of severity for metopic CS. This method may help clinicians easily quantify the severity and perform robust longitudinal assessments of the condition.

Self-Supervised Discovery of Anatomical Shape Landmarks.

Statistical shape analysis is a very useful tool in a wide range of medical and biological applications. However, it typically relies on the ability to produce a relatively small number of features that can capture the relevant variability in a population. State-of-the-art methods for obtaining such anatomical features rely on either extensive preprocessing or segmentation and/or significant tuning and post-processing. These shortcomings limit the widespread use of shape statistics. We propose that effective shape representations should provide sufficient information to align/register images. Using this assumption we propose a self-supervised, neural network approach for automatically positioning and detecting landmarks in images that can be used for subsequent analysis. The network discovers the landmarks corresponding to anatomical shape features that promote good image registration in the context of a particular class of transformations. In addition, we also propose a regularization for the proposed network which allows for a uniform distribution of these discovered landmarks. In this paper, we present a complete framework, which only takes a set of input images and produces landmarks that are immediately usable for statistical shape analysis. We evaluate the performance on a phantom dataset as well as 2D and 3D images.

Scribble-Based 3D Shape Segmentation via Weakly-Supervised Learning.

Shape segmentation is a fundamental problem in shape analysis. Previous research shows that prior knowledge helps to improve the segmentation accuracy and quality. However, completely labeling each 3D shape in a large training data set requires a heavy manual workload. In this paper, we propose a novel weakly-supervised algorithm for segmenting 3D shapes using deep learning. Our method jointly propagates information from scribbles to unlabeled faces and learns deep neural network parameters. Therefore, it does not rely on completely labeled training shapes and only needs a really simple and convenient scribble-based partially labeling process, instead of the extremely time-consuming and tedious fully labeling processes. Various experimental results demonstrate the proposed method's superior segmentation performance over the previous unsupervised approaches and comparable segmentation performance to the state-of-the-art fully supervised methods.

A Cooperative Autoencoder for Population-Based Regularization of CNN Image Registration.

Spatial transformations are enablers in a variety of medical image analysis applications that entail aligning images to a common coordinate systems. Population analysis of such transformations is expected to capture the underlying image and shape variations, and hence these transformations are required to produce anatomically feasible correspondences. This is usually enforced through some smoothness-based generic metric or regularization of the deformation field. Alternatively, population-based regularization has been shown to produce anatomically accurate correspondences in cases where anatomically unaware (i.e., data independent) regularization fail. Recently, deep networks have been used to generate spatial transformations in an unsupervised manner, and, once trained, these networks are computationally faster and as accurate as conventional, optimization-based registration methods. However, the deformation fields produced by these networks require smoothness penalties, just as the conventional registration methods, and ignores population-level statistics of the transformations. Here, we propose a novel neural network architecture that simultaneously learns and uses the population-level statistics of the spatial transformations to regularize the neural networks for unsupervised image registration. This regularization is in the form of a bottleneck autoencoder, which learns and adapts to the population of transformations required to align input images by encoding the transformations to a low dimensional manifold. The proposed architecture produces deformation fields that describe the population-level features and associated correspondences in an anatomically relevant manner and are statistically compact relative to the state-of-the-art approaches while maintaining computational efficiency. We demonstrate the efficacy of the proposed architecture on synthetic data sets, as well as 2D and 3D medical data.

Detecting 3D Points of Interest Using Multiple Features and Stacked Auto-encoder.

Considering the fact that points of interest on 3D shapes can be discriminated from a geometric perspective, it is reasonable to map the geometric signature of a point $p$p to a probability value encoding to what degree $p$p is a point of interest, especially for a specific class of 3D shapes. Based on the observation, we propose a three-phase algorithm for learning and predicting points of interest on 3D shapes by using multiple feature descriptors. Our algorithm requires two separate deep neural networks (stacked auto-encoders) to accomplish the task. During the first phase, we predict the membership of the given 3D shape according to a set of geometric descriptors using a deep neural network. After that, we train the other deep neural network to predict a probability distribution defined on the surface representing the possibility of a point being a point of interest. Finally, we use a manifold clustering technique to extract a set of points of interest as the output. Experimental results show superior detection performance of the proposed method over the previous state-of-the-art approaches.

Average Vector Field Integration for St. Venant-Kirchhoff Deformable Models.

We propose Average Vector Field (AVF) integration for simulation of deformable solids in physics-based animation. Our method achieves exact energy conservation for the St. Venant-Kirchhoff material without any correction steps or extra parameters. Exact energy conservation implies that our resulting animations 1) cannot explode and 2) do not suffer from numerical damping, which are two common problems with previous numerical integration techniques. Our method produces lively motion even with large time steps as typically used in physics-based animation. Our implicit update rules can be formulated as a minimization problem and solved in a similar way as optimization-based backward Euler, with only a mild computing overhead. Our approach also supports damping and collision response models, making it easy to deploy in practical computer animation pipelines.

Transparent rutile TiO2 films prepared by thermal oxidation of sputtered Ti on FTO glass.

TiO2 films were prepared via a two-step fabrication process, i.e. deposition of Ti films by magnetron sputtering on an FTO glass substrate followed by thermal oxidation at 600-725 °C. The investigated parameters were Ti layer thickness, temperature of oxidation and deposition conditions (pre-treatment and substrate heating). Such TiO2 films have a rutile structure and contain metallic Sn which is the result of a thermal reaction at the interface between SnO2 and Ti at temperatures above 500 °C. A calcination temperature of 600 °C is optimal for fabricating TiO2 films with significant photoelectrochemical response. Heating of the FTO substrate during magnetron sputtering deposition of Ti films results in a significant improvement of the compactness of the TiO2 films. A similar but not so pronounced improvement was observed for the TiO2 films deposited on the FTO substrate pre-treated with radio-frequency plasma under Ar-O2 and N2-H2 atmosphere. The observed correlation between the increased content of Sn in the TiO2 films and compactness of the TiO2 films supports the explanation of both positive effects by better adhesion of the Ti films to the FTO substrate.

Nanocrystalline Boron-Doped Diamond as a Corrosion-Resistant Anode for Water Oxidation via Si Photoelectrodes.

Due to its high sensitivity to corrosion, the use of Si in direct photoelectrochemical (PEC) water-splitting systems that convert solar energy into chemical fuels has been greatly limited. Therefore, the development of low-cost materials resistant to corrosion under oxidizing conditions is an important goal toward a suitable protection of otherwise unstable semiconductors used in PEC cells. Here, we report on the development of a protective coating based on thin and electrically conductive nanocrystalline boron-doped diamond (BDD) layers. We found that  BDD layers protect the underlying Si photoelectrodes over a wide pH range (1-14) in aqueous electrolyte solutions. A BDD layer maintains an efficient charge carrier transfer from the underlying silicon to the electrolyte solution. Si|BDD photoelectrodes show no sign of performance degradation after a continuous PEC treatment in neutral, acidic, and basic electrolytes. The deposition of a cobalt phosphate (CoPi) oxygen evolution catalyst onto the BDD layer significantly reduces the overpotential for water oxidation, demonstrating the ability of  BDD layers to substitute the transparent conductive oxide coatings, such as indium tin oxide (ITO) and fluorine-doped tin oxide (FTO), frequently used as protective layers in Si photoelectrodes.