Single 2D Research Articles

Density functional theory–guided electrochemical determination of nitrite in foods based on 2D metal-organic frameworks

This paper presents results for four types of morphology-controlled two-dimensional cobalt based tetrakis (4-carboxyphenyl) porphyrin metal-organic frameworks (2D Co-TCPP MOFs) with different catalytic properties for nitrite, and an electrochemical sensor was fabricated by the Co-TCPP MOF with superior catalytic properties. The sensor exhibited a high sensitivity of 311.3 μg/g, an extra low detection limit (LOD) of 0.03 μM, and a wide linear range of 0.1–375 μM, which was applied successfully for nitrite detection in water, milk and sausage. Moreover, density functional theory (DFT) calculations were used to establish the sensing method, the adsorption structures of nitrite on the surface of Co-TCPP MOF were built and the electrocatalytic mechanism was confirmed. The results explained why the sensor fabricated by a single Co-TCPP MOF exhibited equal or better performance than reported sensors fabricated by composite materials based on a combination of theory and experiment. This work reported a universal electrochemical sensing method for nitrite detection in food based on a single morphology-controlled 2D TCPP MOF, which provided a new and valuable strategy for establishing an efficient food safety determination method guided by theoretical calculations.

Three-dimensional shape generation via variational autoencoder generative adversarial network with signed distance function

<span lang="EN-US">Mesh-based 3-dimensional (3D) shape generation from a 2-dimensional (2D) image using a convolution neural network (CNN) framework is an open problem in the computer graphics and vision domains. Most existing CNN-based frameworks lack robust algorithms that can scale well without combining different shape parts. Also, most CNN-based algorithms lack suitable 3D data representations that can fit into CNN without modification(s) to produce high-quality 3D shapes. This paper presents an approach that integrates a variational autoencoder (VAE) and a generative adversarial network (GAN) called 3 dimensional variational autoencoder signed distance function generative adversarial network (3D-VAE-SDFGAN) to create a 3D shape from a 2D image that considerably improves scalability and visual quality. The proposed method only feeds a single 2D image into the network to produce a mesh-based 3D shape. The network encodes a 2D image of the 3D object into the latent representations, and implicit surface representations of 3D objects corresponding to those 2D images are subsequently generated. Hence, a signed distance function (SDF) is proposed to maintain object inside-outside information in the implicit surface representation. Polygon mesh surfaces are then produced using the marching cubes algorithm. The ShapeNet dataset was used in the experiments to evaluate the proposed 3D-VAE-SDFGAN. The experimental results show that 3D-VAE-SDFGAN outperforms other state-of-the-art models.</span>

Application of photogrammetry and laser scanner on the Bronze Age structures of the Castillejos de Luna cist tomb necropolis (Pizarra, Spain)

The cist tombs necropolis at Castillejos de Luna, in Sierra de Gibralmora-Sierra del Hacho (Pizarra, Málaga, Spain) was known from the graphic documentation and the grave goods of two tombs. New studies have documented nine burials. The aim of this article is to present the new virtualisation work that has been carried out in the necropolis, to generate a new three-dimensional (3D) documentation of the currently known records, which are in acceptable visibility conditions. Using tools to document tombs in 3D offers us great advances in data acquisition and editing, with great precision and realism, thanks to the 3D models generated through techniques such as photogrammetry or laser scanning. Thanks to these tools, it is possible to carry out studies on digital twins and use them as informative material for society. The study this paper describes has generated high quality products for dissemination and future analysis. The results shown here are of metric character, with orientation and geographical location of the structures. In addition, in one of the tombs the authors show the combination of photogrammetric techniques with laser scanners to obtain a single high-resolution 3D model; subsequently a retopology process is carried out to achieve a "light" model with a photorealistic appearance that is both easily manipulated on mobile devices for its dissemination and a guarantee that the general public can enjoy this necropolis in a different way. The preliminary results are published in the web repository of 3D models Sketchfab, where the users can see a preview of one of the tombs before and after being optimised with retopology through Blender. The authors provide a socio-historical analysis of Bronze Age necropolises in central Andalusia, within the framework of a debate on the western expansion of the El Argar Culture.

A cyber-physical system to design 3D models using mixed reality technologies and deep learning for additive manufacturing.

I-nteract is a cyber-physical system that enables real-time interaction with both virtual and real artifacts to design 3D models for additive manufacturing by leveraging mixed-reality technologies. This paper presents novel advances in the development of the interaction platform to generate 3D models using both constructive solid geometry and artificial intelligence. In specific, by taking advantage of the generative capabilities of deep neural networks, the system has been automated to generate 3D models inferred from a single 2D image captured by the user. Furthermore, a novel generative neural architecture, SliceGen, has been proposed and integrated with the system to overcome the limitation of single-type genus 3D model generation imposed by differentiable-rendering-based deep neural architectures. The system also enables the user to adjust the dimensions of the 3D models with respect to their physical workspace. The effectiveness of the system is demonstrated by generating 3D models of furniture (e.g., chairs and tables) and fitting them into the physical space in a mixed reality environment. The presented developmental advances provide a novel and immersive form of interaction to facilitate the inclusion of a consumer into the design process for personal fabrication.

Auto-segmentation of thoracic organs in CT scans of breast cancer patients using a 3D U-net cascaded into 2D patchGANs

The quality of organ volume delineation significantly influences the efficacy of radiotherapy treatment for breast cancer patients. This study introduces a novel method for auto-segmentation of the breasts, lungs and heart. The proposed pipeline leverages a multi-class 3D U-Net with a pre-trained ResNet(2+1)D-18 encoder branch, cascaded with a 2D PatchGAN mask correction model for each class. This approach requires a single 3D model, providing a relatively efficient solution. The models were trained and evaluated on 70 thoracic DICOM datasets belonging to breast cancer patients. The evaluation demonstrated state-of-the-art segmentation performance, with mean Dice similarity coefficient values ranging from 0.89 to 0.98, Hausdorff distance values ranging from 2.25 to 8.68 mm, and mean surface distance values ranging from 0.62 to 2.79 mm. These results underscore the pipeline’s potential to enhance breast cancer diagnosis and treatment strategies, with possible applications in other medical sectors utilizing auto-segmentation.

High Carrier Mobility and Controllable Electronic Property of the h-BN/SnSe2 Heterostructure.

Building two-dimensional (2D) vertical van der Waals heterostructures (vdWHs) is one of the effective methods to regulate the properties of single 2D materials. In this paper, we stack the hexagonal boron nitride (h-BN) monolayer (ML) on the SnSe2 ML to construct the stable h-BN/SnSe2 vdWH, of which the crystal and electronic structures, together with the optical properties, are also analyzed by the first-principles calculations. The results show that the h-BN/SnSe2 vdWH belongs to a type-I heterostructure with an indirect bandgap of 1.33 eV, in which the valence band maximum and conduction band minimum are both determined by the component SnSe2 ML. Interestingly, the h-BN/SnSe2 vdWH under the tensile strain or electric field undergoes the transitions both from type-I to type-II heterostructure and from the indirect to direct bandgap semiconductor. In addition, the carrier mobility of the h-BN/SnSe2 heterostructure has a significant enhancement relative to that of the SnSe2 ML, up to 104 cm2 V-1 s-1. Meanwhile, the h-BN/SnSe2 heterostructure presents the superb optical absorption and unique type-II hyperbolic property. Our findings will broaden the potential applications of SnSe2 ML and provide theoretical guidance for the related experimental studies.

3DCascade-GAN: Shape completion from single-view depth images

Depth images can be easily acquired using depth cameras. However, these images only contain partial information about the shape due to unavoidable self-occlusion. Thanks to the availability of large datasets of shapes, it is possible to use a learning-based approach to produce complete shapes from single depth images. State-of-the-art generative adversarial network (GAN) architectures can produce reasonable results. However, the use of relatively local convolutions restricts GAN architectures from producing globally plausible shapes. In this study, we develop a novel dynamic latent code selection mechanism in which the model learns to select only important codes from the latent space. Furthermore, a novel 3D self-attention (3DSA) layer is introduced that is able to capture non-local relationships across the 3D space. We further design a GAN architecture that uses a multistage encoder–decoder to recover the shape, where our 3DSA layer is introduced to the discriminator to help attend to global features, which stabilizes the model learning and encourages shape refinement, making our reconstruction more structurally plausible. Through extensive experiments, we demonstrate that our method outperforms other state-of-the-art methods for single depth image 3D reconstruction.

Single exposure passive three-dimensional information reconstruction based on an ordinary imaging system

Existing three-dimensional (3D) imaging technologies have issues such as requiring active illumination, multiple exposures, or coding modulation. We propose a passive single 3D imaging method based on an ordinary imaging system. Using the point spread function of the imaging system to realize the non-coding measurement on the target, the full-focus images and depth information of the 3D target can be extracted from a single two-dimensional (2D) image through the compressed sensing algorithm. Simulation and experiments show that this approach can complete passive 3D imaging based on an ordinary imaging system without any coding operations. This method can achieve millimeter-level vertical resolution under single exposure conditions and has the potential for real-time dynamic 3D imaging. It improves the efficiency of 3D information detection, reduces the complexity of the imaging system, and may be of considerable value to the field of computer vision and other related applications.

Robot for thermal monitoring of buildings

Until the last decade, robots for thermal monitoring of buildings were limited to measuring ambient temperature. Recent research has focused on obtaining detailed thermal information about the structural components of buildings since the mid-2010s. This paper presents the methodology and results of the MoPAD2 (Mobile Platform for Autonomous Digitization) robot, an autonomous, low-cost, small-sized mobile robot equipped with a thermal scanning platform. The platform generates 3D thermal models of building interiors over time and conducts repeated data acquisition sessions, saving energy during waiting intervals. The paper explains how a dense thermal point cloud is obtained from different locations in the scene and processed to create a single 3D thermal model, and two case studies are presented. This robotics system has the potential to revolutionize thermal monitoring of buildings and enable energy characterization through thermal models.

UV polymerization fabrication method for polymer composite based optical fiber sensors

Optical fiber (OF) sensors are critical optical devices with excellent sensing capabilities and the capacity to operate in remote and hostile environments. However, integrating functional materials and micro/nanostructures into the optical fiber systems for specific sensing applications has limitations of compatibility, readiness, poor control, robustness, and cost-effectiveness. Herein, we have demonstrated the fabrication and integration of stimuli-responsive optical fiber probe sensors using a novel, low-cost, and facile 3D printing process. Thermal stimulus–response of thermochromic pigment micro-powders was integrated with optical fibers by incorporating them into ultraviolet-sensitive transparent polymer resins and then printed via a single droplet 3D printing process. Hence, the thermally active polymer composite fibers were grown (additively manufactured) on top of the commercial optical fiber tips. Then, the thermal response was studied within the temperature range of (25–35 °C) and (25–31 °C) for unicolor and dual color pigment powders-based fiber-tip sensors, respectively. The unicolor (with color to colorless transition) and dual color (with color to color transition) powders-based sensors exhibited substantial variations in transmission and reflection spectra by reversibly increasing and decreasing temperatures. The sensitivities were calculated from the transmission spectra where average change in transmission spectra was recorded as 3.5% with every 1 °C for blue, 3% for red and 1% for orange-yellow thermochromic powders based optical fiber tip sensors. Our fabricated sensors are cost-effective, reusable, and flexible in terms of materials and process parameters. Thus, the fabrication process can potentially develop transparent and tunable thermochromic sensors for remote sensing with a much simpler manufacturing process compared to conventional and other 3D printing processes for optical fiber sensors. Moreover, this process can integrate micro/nanostructures as patterns on the optical fiber tips to increase sensitivity. The developed sensors may be employed as remote temperature sensors in biomedical and healthcare applications.

EfficientNetV2-based dynamic gesture recognition using transformed scalogram from triaxial acceleration signal

Abstract In this paper, a dynamic gesture recognition system is proposed using triaxial acceleration signal and image-based deep neural network. With our dexterous glove device, 1D acceleration signal can be measured from each finger and decomposed to time-divided frequency components via wavelet transformation, which is known as scalogram as image-like format. To feed-forward the scalogram with single 2D, convolutional neural networks allows the gesture having temporality to be easily recognized without any complex system such as RNN, LSTM, or spatio-temporal feature as 3D CNN, etc. To classify the image with general input dimension of image RGB channels, we numerically reconstruct fifteen scalograms into one RGB image with various representation methods. In experiments, we employ the off-the-shelf model, EfficientNetV2 small-to-large model as an image classification model with fine-tuning. To evaluate our system, we bulid our custom bicycle hand signals as dynamic gesture dataset under our transformation system, and then qualitatively compare the reconstruction method with matrix representation methods. In addition, we use other signal transformation tools such as the fast Fourier transform and short-time Fourier transform and then explain the advantages of scalogram classification in the terms of time-frequency resolution trade-off issue.

Femoral coordinate system based on articular surfaces: Implications for computer-assisted knee arthroplasty

Osteoarthritis knee can be restored by total knee arthroplasty (TKA). Imageless TKA requires several anatomical points to construct a reference coordinate system to measure bone resections and implant placement. Inaccuracies in the definition of the coordinate system lead to malalignment and failure of the implant. While the surgical transepicondylar axis (sTEA) is a reliable anatomical axis to define the lateromedial axis for the femoral coordinate system (FCS), the presence of the collateral ligaments and deterioration of the medial sulcus (MS) make the registration of sTEA a challenging task. In this work, sTEA is assigned using the articular surfaces of the femoral condyles, independent of the lateral epicondyle (LE) and MS. A single 3D arc is marked on each condyle, which is transformed into a 2D arc to get the best-fit curve according to the profile of condyles. The turning point of each best-fit curve, when transformed back to 3D, defines an axis parallel to sTEA. The condyles-based sTEA is measured experimentally on a 3D-printed bone using an Optitrack tracking setup. Using the proposed method, the angle between the aTEA, sTEA, and Whiteside's line was (3.77, 0.55, and 92.72)°, respectively. The proposed method provides the same level of accuracy and improves the anatomical points registration efficiency, as there is no need to register the LE or MS.

Synthesis of Single Crystal 2D Cu2FeSnS4 Nanosheets with High-Energy Facets (111) as a Pt-Free Counter Electrode for Dye-Sensitized Solar Cells.

Two-dimensional Cu2FeSnS4 (CFTS) nanosheets with exposed high-energy facets (111) have been synthesized by a facile, scalable, and cost-effective one-pot heating process. The CFTS phase formation is confirmed by both X-ray diffraction and Raman spectroscopy. The formation mechanism of exposed high-energy facet CFTS growth is proposed and its electrochemical and photoelectrochemical properties are investigated in detail to reveal the origin of the anisotropic effect of the high-energy facets. Dye-sensitized solar cells (DSSC) achieve a favorable power conversion efficiency of 5.92% when employing CFTS thin film as a counter electrode, suggesting its potential as a cost-effective substitute for Pt in DSSCs.

Computationally efficient GPU based NS solver for two dimensional high-speed inviscid and viscous compressible flows

In this study, we proposed a novel GPU-based solution for modelling two-dimensional inviscid and viscous compressible supersonic/hypersonic flows. Texture and surface pointers are used to access GPU memory locations. For effective and efficient use of surface pointers, we grouped multiple 2D arrays referenced and indexed by a single 3D surface pointer. To enable the proposed solver for double-precision calculations, two consecutive 32-bit memory locations were grouped to maintain the efficiency of surface pointers while taking advantage/accuracy of 64-bit calculations. Resolving data and computation dependencies for parallel applications is another complex task that is the focus of this study. Computation dependencies have been solved by using multiple mutually synchronized GPU kernels and executing them sequentially using the GPU default stream to ensure that all relevant data is available to the threads or computed before they actually use it. Consequently, there is no intra-core data dependency in our proposed approach, while inter-core data dependency is successfully solved by stringing multiple kernels together. Using NVIDIA GTX 660 GPUs, we achieved 20x speedup compared to traditional Core i5® computers. This speedup is the result of the Surface Pointer's GPU capabilities for double precision computations. The simulation results are also consistent with the experimental and numerical results of this study.

Localization and Registration of 2D Histological Mouse Brain Images in 3D Atlas Space.

To accurately explore the anatomical organization of neural circuits in the brain, it is crucial to map the experimental brain data onto a standardized system of coordinates. Studying 2D histological mouse brain slices remains the standard procedure in many laboratories. Mapping these 2D brain slices is challenging; due to deformations, artifacts, and tilted angles introduced during the standard preparation and slicing process. In addition, analysis of experimental mouse brain slices can be highly dependent on the level of expertise of the human operator. Here we propose a computational tool for Accurate Mouse Brain Image Analysis (AMBIA), to map 2D mouse brain slices on the 3D brain model with minimal human intervention. AMBIA has a modular design that comprises a localization module and a registration module. The localization module is a deep learning-based pipeline that localizes a single 2D slice in the 3D Allen Brain Atlas and generates a corresponding atlas plane. The registration module is built upon the Ardent python package that performs deformable 2D registration between the brain slice to its corresponding atlas. By comparing AMBIA's performance in localization and registration to human ratings, we demonstrate that it performs at a human expert level. AMBIA provides an intuitive and highly efficient way for accurate registration of experimental 2D mouse brain images to 3D digital mouse brain atlas. Our tool provides a graphical user interface and it is designed to be used by researchers with minimal programming knowledge.

Robot Motion Planning with Optimization-based Algorithm and Neural Network: Unsupervised Path Regression

Planning collision-free motions for robots is challenging in environments with complex obstacle geometries. In this study, we combine the optimization-based algorithm and neural network to obtain a short and collision-free path through unsupervised path regression. The network can help to predict an educated initial path for an optimization-based planner, which will converge to a better solution. We compare different path representations, world representations, and also their combinations to improve the result. The motion planning problem is extended from a sphere robot in a single 2D world to a more complex static arm robot in multiple worlds. Through our experiments, we can find that in multiple worlds, the relative path with the signed distance field is the best combination for both robots. This is not only proved by metrics of feasible rate and loss but also by the visualization in the map.

FacT: Factor-Tuning for Lightweight Adaptation on Vision Transformer

Recent work has explored the potential to adapt a pre-trained vision transformer (ViT) by updating only a few parameters so as to improve storage efficiency, called parameter-efficient transfer learning (PETL). Current PETL methods have shown that by tuning only 0.5% of the parameters, ViT can be adapted to downstream tasks with even better performance than full fine-tuning. In this paper, we aim to further promote the efficiency of PETL to meet the extreme storage constraint in real-world applications. To this end, we propose a tensorization-decomposition framework to store the weight increments, in which the weights of each ViT are tensorized into a single 3D tensor, and their increments are then decomposed into lightweight factors. In the fine-tuning process, only the factors need to be updated and stored, termed Factor-Tuning (FacT). On VTAB-1K benchmark, our method performs on par with NOAH, the state-of-the-art PETL method, while being 5x more parameter-efficient. We also present a tiny version that only uses 8K (0.01% of ViT's parameters) trainable parameters but outperforms full fine-tuning and many other PETL methods such as VPT and BitFit. In few-shot settings, FacT also beats all PETL baselines using the fewest parameters, demonstrating its strong capability in the low-data regime.

Image-based 3D modeling-to-simulation of single-wythe masonry structure via reverse descriptive geometry

This study aims to develop an innovative image-based 3D modeling-to-simulation framework that can efficiently assess the hazard vulnerability of single-wythe masonry structures. Using the novel concept of ‘reverse descriptive geometry’, the framework utilizes 2D masonry wall images to model a 3D single-wythe masonry structure, subsequently converting it into a 3D discrete element model for the vulnerability assessment. The methodology involves segmenting all bricks in the 2D images, identifying the brick geometries by tracing the boundaries, and approximating those into polygons using the splitting method. By incorporating the wall orientations, the simplified 2D polygons are introduced into 3D space and subsequently extruded to a specified thickness. The corners between the walls, i.e., adjoining wythes, are realistically modeled with a type of structural masonry bond. The digitally constructed 3D masonry geometry is enriched with physical properties such as brick mass and mortar strengths, enabling a seamless 3D discrete element simulation to gauge the hazard vulnerability. The proposed framework is successfully applied to Stylite Tower, a heritage single-wythe masonry structure in Jordan. The results demonstrate a remarkable geometric accuracy achieved and further illustrate the structure's potential vulnerability under seismic events, highlighting the efficiency and efficacy of the developed framework. This study introduces a streamlined procedure leveraging a single 2D image per wall for 3D modeling of a structure, thus highly efficient. A discrete element modeling is adopted instead of finite element methods based on a continuum theory. Therefore, this approach accurately captures the discontinuous nature of masonry structures, resulting in a high fidelity in simulation. Furthermore, this study adopts an impulse-based dynamic simulation method that is far more computationally efficient and tolerant to geometric errors than the conventional discrete element method. Therefore, the proposed framework presents a significant advancement in the vulnerability assessment of masonry structures, providing a novel, efficient, and reliable methodology for engineering practice.

1747-P: Islet Ca2+ Dynamics, Heterogeneity, and Consistency in Three Dimensions with Activators of Pyruvate Kinase

Pancreatic β-cell heterogeneity is important for Ca2+ triggering and insulin secretion and is affected in diabetes. Organization and consistency of β-cell subpopulations are poorly understood despite being key in understanding how individual cells contribute to islet function and diabetes. Existing studies analyze single 2D planes of the islet. However, subpopulation organization cannot be fully understood without examining β cell dynamics throughout the entire islet in 3D. Here, we imaged islets from β cell specific GCaMP6s Ca2+ reporter mice using high-speed light-sheet microscopy to analyze the 3D organization and consistency in high glucose and with pharmacologic activators of pyruvate kinase (PK), a glycolytic enzyme that controls βcell excitability. Phase analysis revealed that cells which initiated islet depolarization (leaders) are not highly consistent across oscillations, but the 3D region where the calcium wave originates is consistent, according to a metric based upon the KL-divergence (P<0.0001 over 11 oscillations). Using functional network analysis, β-cell hubs which drive islet synchrony are more consistent over an islet region than across individual cells (P<0.0001). The wave origin was more regionally dependent than the β-cell hubs (P<0.0001). We also found that network analysis is similar (P=0.04) and phase analysis is different between 2D and 3D (P<0.0001). Following quantification of heterogeneity in the 3D islet, certain sub-populations of β cells are similar throughout the islet, while others are not. Acute PK activation decreased the consistency of wave initiators (P<0.0001), but not the consistency of hub cells in the functional network (P=0.15), shedding new light onto the mechanism behind these subpopulations. Islet regions, rather than individual cells, are more consistent and thus likely responsible for initiating Ca2+ triggering and insulin secretion. These mechanisms help render the islet robust to single β cell failures. Disclosure J.Briggs: None. E.Jin: None. M.J.Merrins: None. R.K.Benninger: None. Funding William S. Middleton Memorial Veterans Hospital; U.S. Department of Veterans Affairs Biomedical Laboratory Research and Development Service (I01B005113); National Institutes of Health (R01DK113103, R01DK127637, R01DK102950, R01DK106412); National Science

A semantic 3D city model for underground land administration: Development and implementation of an ADE for CityGML 3.0

Current practices of managing and communicating underground physical and cadastral (legal) data are based on fragmented and silo-based 2D methods, which cannot provide unambiguous, coordinated, and reliable representations of assets and their rights, restrictions, and responsibilities (RRRs) in subterranean environments. There is also no link between underground physical and legal data in current practices, resulting in ineffective Underground Land Administration (ULA). A single 3D integrated model of underground assets associated with the cadastral data at a city-scale can improve understanding and interpretation of underground legal and physical aspects. Extending a 3D city model to include underground legal data requirements can potentially provide a viable approach for developing a 3D integrated model. This research aims to develop a new Application Domain Extensions (ADE) for CityGML 3.0 (called VicULA ADE) to manage underground data using a single 3D integrated legal-physical model at a city-scale. The VicULA ADE proposes semantic entities for underground legal spaces, enabling the model to define spatial and semantic relationships between different types of legal spaces. The proposed approach comprises identifying requirements based on the Victorian state of Australia, developing the conceptual and logical data models in UML, and implementing the XML schema as the physical data model. The viability of the VicULA ADE is tested in a complex real-world case study that has various underground legal and physical objects, shifting this data model from a conceptual level to the implementation level. The proposed ADE of the CityGML 3.0 standard revealed how underground legal data elements can be logically embedded in this 3D data model.