Search

Meshes obtained by 3D scanning or photogrammetry are usually very dense and multi-chart textured. Effectively reducing the amount of data while preserving both geometry and appearance is an important and also challenging problem. This paper presents a new method for seamless simplification of dense triangular meshes with multi-chart textures. The method is composed of three components: geometry simplification, correspondence map construction and texture image generation. The core of the method lies in the ideas of decoupling geometry, UV-parameterization and texture processes to make the simplification process less constrained by the underlying attribute encoding, a reliable updating procedure for constructing the correspondence map between the simplified and original meshes, and a least squares model for filling texture elements, which are the key to producing artifact-free visual appearance. Compared to the state-of-the-art, the proposed method has several advantages: (1) it enables aggressive simplification with well preserved geometry and appearance; (2) it automatically achieves seamless simplification without special treatment on the seams between multi-charts of the texture; and (3) it is reliable. Experimental results have demonstrated these features. Nanyang Technological University National Research Foundation (NRF) This work was supported by a joint WASP/NTU project (04INS000440C130), Ynnerman Scholar (KAW 2019.0024) and National Research Foundation (NRF) Singapore, under its Virtual Singapore (VSG) Programme (Award No. NRF2015VSG-AA3DCM001-018).

Micronutrient malnutrition is one of the most common public health problems in the world. Biofortification the most attractive and sustainable solution to surmount malnutrition requires the development of micronutrient enriched new crop cultivars. In this study, two RIL populations, ZM175/XY60 and ZM175/LX987, were used to identify QTL for grain zinc concentration (GZnC), grain iron concentration (GFeC) and thousand grain weight (TGW). Eight QTL for GZnC, six QTL for GFeC and five QTL for TGW were detected. Three QTL on chromosome arms 2DL and 4BS and chromosome 6A showed pleiotropic effects on all three traits. The 4BS and 6A QTL also increased plant height and might be Rht-B1a and Rht25a, respectively. The 2DL locus within a suppressed recombination region was identified in both RIL populations and the favorable allele simultaneously increasing GZnC, GFeC and TGW was contributed by XY60 and LX987. A QTL on chromosome arm 6DL associated only with GZnC was detected in ZM175/XY60 and was validated in a JD8/AK58 RILs using KASP marker K_AX-110119937. Both the 2DL and 6DL QTL were new loci for GZnC. Based on gene annotations, sequence variations and expression profiles, phytic acid biosynthesis gene TaIPK1-2D and nicotianamine synthase gene TaNAS10-6D were predicted as candidate genes. Their gene-based KASP markers were developed and were validated in a cultivar panel of 343 wheat accessions. This study investigated the genetic basis of GZnC and GFeC, and provided valuable candidate genes and markers for breeding Zn- and Fe-enriched wheat.

We present a simple yet effective method for constructing 3D self-supporting surfaces with planar quadrilateral (PQ) elements. Starting with a triangular discretization of a self-supporting surface, we first compute the principal curvatures and directions of each triangular face using a new discrete differential geometry approach, yielding more accurate results than existing methods. Then, we smooth the principal direction field to reduce the number of singularities. Next, we partition all faces into two groups in terms of principal curvature difference. For each face with small curvature difference, we compute a stretch matrix that turns the principal directions into a pair of conjugate directions. For the remaining triangular faces, we simply keep their smoothed principal directions. Finally, applying a mixed-integer programming solver to the mixed principal and conjugate direction field, we obtain a planar quadrilateral mesh. Experimental results show that our method is computationally efficient and can yield high-quality PQ meshes that well approximate the geometry of the input surfaces and maintain their self-supporting properties. Ministry of Education (MOE) Published version This work was partially supported by National Natural Science Foundation of China (62172257, 61802228), Singapore Ministry of Education (T2EP20220-0014), and the RIE2020 Industry Alignment Fund–Industry Collaboration Projects (IAF–ICP) Funding Initiative, as well as cash and in-kind contribution from the industrial partner, Rolls-Royce.

Many real-world problems, such as airfoil design, involve optimizing a black-box expensive objective function over complex-structured input space (e.g., discrete space or non-Euclidean space). By mapping the complex-structured input space into a latent space of dozens of variables, a two-stage procedure labeled as generative model-based optimization (GMO), in this article, shows promise in solving such problems. However, the latent dimension of GMO is hard to determine, which may trigger the conflicting issue between desirable solution accuracy and convergence rate. To address the above issue, we propose a multiform GMO approach, namely, generative multiform optimization (GMFoO), which conducts optimization over multiple latent spaces simultaneously to complement each other. More specifically, we devise a generative model which promotes a positive correlation between latent spaces to facilitate effective knowledge transfer in GMFoO. And furthermore, by using Bayesian optimization (BO) as the optimizer, we propose two strategies to exchange information between these latent spaces continuously. Experimental results are presented on airfoil and corbel design problems and an area maximization problem as well to demonstrate that our proposed GMFoO converges to better designs on a limited computational budget.

Rye (Secale cereale L.) is an important genetic resource for improving the disease resistance of wheat. An increasing number of rye chromosome segments have been transferred into modern wheat cultivars via chromatin insertions. In this study, 185 recombinant inbred lines (RILs) derived from a cross between a common wheat accession containing rye chromosomes 1RS and 3R and a wheat-breeding founder parent (Chuanmai 42) from southwestern China were used to decipher the cytological and genetic effects of 1RS and 3R via fluorescence/genomic in situ hybridization and quantitative trait locus (QTL) analyses. Chromosome breakage was detected in the RIL population. Additionally, the recombination of chromosomes 1BS and 3D from Chuanmai 42 was completely suppressed by 1RS and 3R in the RIL population. In contrast to chromosome 3D of Chuanmai 42, rye chromosome 3R was significantly associated with white seed coats and decreased yield-related traits, as revealed by QTL and single marker analyses, whereas it had no effect on stripe rust resistance. Rye chromosome 1RS did not affect yield-related traits and it increased the susceptibility of plants to stripe rust. Most of the detected QTLs that positively affected yield-related traits were from Chuanmai 42. The findings of this study suggest that the negative effects of rye-wheat substitutions or translocations, including the suppression of the pyramiding of favorable QTLs on paired wheat chromosomes from different parents and the transfer of disadvantageous alleles to filial generations, should be considered when selecting germplasm to enhance wheat-breeding founder parents or to breed new varieties.

This paper introduces a special class of 3D Bézier curves that are defined by their degree, a starting point, the first leg of their control polygons, and a 3D affine transformation composing of a uniform scaling and a rotation. We present new formulas for the curvature of such Bézier curves and based on the new formulas we derive sufficient conditions for the curves to have monotonic curvature. The conditions are expressed by a simple constraint on the rotation angle and the scaling factor. This facilitates constructing 3D Class A Bézier curves that are a generalization of planar typical curves proposed by Mineur et al. (1998), which are often used in automotive and other design applications. Some examples are provided to demonstrate the effectiveness of our construction. Ministry of Education (MOE) This work was supported by the Opening Fund of State Key Laboratory of Lunar and Planetary Sciences (Macau FDCT grant No. 119/2017/A3, Macau University of Science and Technology), the Natural Science Foundation of China (Project No. 61572056), and MOE AcRF Tier 1 Grant of Singapore (RG12/22).

Primary central nervous system lymphomas (PCNSLs) often present a unique histopathological feature of aggregative perivascular tumor cells (APVT). Our previous studies showed that patients of PCNSL with APVTs exhibited poor long-term outcomes and increased expression of the endoplasmic reticulum stress (ERS) factor X-box-binding protein (XBP1). However, very little is known about molecular mechanism of the APVT formation in PCNSLs. The aim of this study is to determine if hypoxia-induced ERS is related to the APVT formation in PCNSLs. In this study, cell culture was used to observe the interplay between diffuse large B cell lymphoma (DLBCL) tumor cells and human brain microvascular endothelial cells (HBMECs) in different oxygen conditions. The expression of XBP1, CXCR and CD44 was manipulated by molecular cloning and siRNA technology. Mouse in vivo experiments and clinical studies were conducted to confirm our hypothesis. Our results showed that activated B-cell type-DLBCL cells easily migrated and invaded, and expressed high levels of XBP1 and stromal molecules CXCR4 and CD44 during hypoxia-induced ERS and dithiothreitol unfolded protein response (UPR). The gene upregulation (using overexpression vector) and downregulation (siRNA gene knock-out) in cultured cells and in mouse models further confirmed a close relation of the expression of XBP1, CXCR4, and CD44 with APVT formation, which is coincided with our clinical observation that increased expression of XBP1, CXCR4, and CD44 in the APVT cells in PCNSLs were associated with poor clinical outcomes. The results suggest that hypoxia-induced ERS and UPR might be associated with APVTs formation in PCNSL and its poor clinical outcomes. The results will help us better understand the progression of PCNSL with APVTs feature in daily pathological work and could be valuable for future target treatment of PCNSLs.

Through many field geology observations, the authors find that sandstone and carbonate rocks, which were originally thought to be sedimentary, have magmatic characteristics and should classified as magmatic rather than sedimentary rocks. Combining this information with the principles of Bowen’s reaction series, the authors infer that the Earth has evolved gradually from high temperature to low temperature and, through crystallization differentiation, has experienced the following stages: high melting point metal to ultramafic rock to mafic rock to neutral rock to acidic rock to sandstone to carbonate rock to mudstone to ice. Since fossils that preserve the remains of creatures, such as fossil fish, occur in magmatic rocks rather than sedimentary rocks, the organisms were not aquatic but lived in high-temperature fluids with magmatic characteristics. Therefore, the authors conclude that during the evolution of rocks from high temperature to low temperature, biological evolution from high temperature to low temperature also occurred on the Earth and that carbon-based life did not originate in the ocean but in high-temperature carbon-containing fluids; this is a subversive statement. This understanding contradicts the original interpretation, that is, that fossils formed under normal temperature and pressure conditions. This point explains the Cambrian explosion and later mass extinctions.

Wheat is the crop most widely grown in the world and provides the daily protein and 20% food calories for 4.5 billion people. It is crucial to understand the genetic basis of grain hardness for improving wheat quality. In order to explore the molecular basis of the formation of wheat grain hardness, two wheat cultivars Chuanmai 66 and Shumai 969 with significant hardness difference in southwest wheat region were selected to analyze the proteins differential expression by TMT quantitative proteomics (tandem mass tags) and bioinformatic methods of function and pathway enrichment analysis. A total of 6020 effective proteins were identified and quantified, including 113 differentially expressed proteins (DEPs), of which 69 were up-regulated and 44 were down-regulated in soft wheat Chuanmai 66. These DEPs were enriched into 65 GO terms, including a biological process term, a cellular component term and six molecular function terms at extremely significant level. Based on the enrichment analysis, we suggested that nutrient reservoir activity proteins, enzyme inhibitor proteins and glutathione metabolism proteins might participate in the formation of wheat grain hardness, and grain hardness related proteins might mainly distribute in the extracellular region of cells and had defensive function. According to the phylogenetic analysis, it was inferred that puroindolines and its homologous proteins might be as not only wheat grain storage proteins, but also enzyme inhibitors regulating grain development. This study provides a basis for further exploring the genetic mechanism of wheat grain hardness.

Stripe rust, caused by Puccinia striiformis f. sp. tritici, is a serious disease in bread wheat (Triticum aestivum L.). Identification and use of adult plant resistance (APR) resources are important for stripe rust resistance breeding. Bread wheat line C33 is an exotic germplasm that has shown stable APR to stripe rust for more than 10 years in Sichuan Province of China. Here, 183 recombinant inbred lines (RILs) derived from the cross between C33 and a susceptible line X440 were genotyped with diversity arrays technology (DArT) markers to identify resistance quantitative trait locus (QTL). Field trials were conducted in five years at Chengdu and Xindu of Sichuan Province, using maximum disease severity (MDS) as stripe rust reaction phenotypes. A total of four quantitative trait loci (QTLs) were detected, respectively designed as QYr.saas-3AS, QYr.saas-5AL, QYr.saas-5BL, and QYr.saas-7DS, explaining 4.14–15.21% of the phenotypic variances. QYr.saas-5BL and QYr.saas-7DS were contributed by C33. However, the level for stripe rust resistance contributed by them was not strong as C33, suggesting the presence of other unidentified QTLs in C33. QYr.saas-7DS corresponded to Yr18 and QYr.saas-5BL remains to be formally named. The RIL lines carrying combinations QYr.saas-5AL, QYr.saas-5BL, and QYr.saas-7DS showed comparability resistance with C33. The present study provides resources to pyramid diverse genes into locally adapted elite germplasm to improve the stripe rust resistance of bread wheat.

This paper presents a method for generative design of decorative architectural parts such as corbel, moulding and panel, which usually have clear structure and aesthetic details. The method is composed of two components: offline learning and online generation. The offline learning trains a 2D CurveInfoGAN and a 3D VoxelVAE that learn the feature representations of the parts in a dataset. The online generation proceeds with an evolution procedure that evolves to product new generation of part components by selecting, crossing over and mutating features, followed by a feature-driven deformation that synthesizes the 3D mesh representation of new models. Built upon these technical components, a generative design tool is developed, which allows the user to input a decorative architectural model as a reference and then generates a set of new models that are “more of the same” as the reference and meanwhile exhibit some “surprising” elements. The experiments demonstrate the effectiveness of the method and also showcase the use of classic geometric modelling and advanced machine learning techniques in modelling of architectural parts. Ministry of Education (MOE) This work is supported by the Ministry of Education, Singapore, under its MoE Tier-2 Grant (2017-T2-1-076).

Octree is a hierarchical data structure with many applications, especially in encoding unstructured point clouds. The depth of an octree is dependent of the scale of the input data and the desired resolution of the smallest voxels in the leaf nodes as well. Thus, it often requires a deep octree to maintain low level of geometric errors for large-scale sparse point clouds, which leads to high memory requirement and low access speed. This paper presents a new structure called truncated octree or T-Octree that truncates the octree by adaptively pruning the top hierarchy and represents the deep octree by a set of shallow sub-octrees. The structure is further extended to support random access of nodes and out-of-core streaming of large data sets. We also propose a variable length addressing scheme to adaptively choose the length of an octree’s node address based on the truncation level. As a result, T-Octree provides highly efficient query performance and can save storage without losing the original structure for sparse or clustered models and scenes. We demonstrate the efficacy and efficiency of the new structure on point cloud compression and scene query tasks for sparse or clustered data. Ministry of Education (MOE) National Research Foundation (NRF) This work was supported by National Research Foundation (NRF) Singapore, under its Virtual Singapore (VSG) Programme (Award No. NRF2015VSG-AA3DCM001-018), and by the Ministry of Education, Singapore, under its MoE Tier-2 Grant (MoE 2017-T2-1-076).

Biofortification is a sustainable strategy to alleviate micronutrient deficiency in humans. It is necessary to improve grain zinc (GZnC) and iron concentrations (GFeC) in wheat based on genetic knowledge. However, the precise dissection of the genetic architecture underlying GZnC and GFeC remains challenging. In this study, high-resolution genome-wide association studies were conducted for GZnC and GFeC by three different models using 166 wheat cultivars and 373,106 polymorphic markers from the wheat 660K and 90K single nucleotide polymorphism (SNP) arrays. Totally, 25 and 16 stable loci were detected for GZnC and GFeC, respectively. Among them, 17 loci for GZnC and 8 for GFeC are likely to be new quantitative trait locus/loci (QTL). Based on gene annotations and expression profiles, 28 promising candidate genes were identified for Zn/Fe uptake (8), transport (11), storage (3), and regulations (6). Of them, 11 genes were putative wheat orthologs of known

Objectives Dynamic contrast-enhanced MRI (DCE-MRI) with Extended Tofts Linear (ETL) model has been used in tissue and tumor evaluation. However, its reliability and reproducibility in pancreatic evaluation has been unclear. It is also unclear whether pancreatic DCE-MRI pharmacokinetic parameters were consistent and stable among different pancreatic regions, ages and genders. Methods Pancreatic pharmacokinetic parameters of 54 volunteers were calculated using DCE-MRI with ETL model. Firstly, Intra- and inter-observer reproducibility was evaluated using intra-class correlation coefficient (ICC) and coefficient of variation (CoV). Secondly, subgroup evaluation of pancreatic DCE-MRI pharmacokinetic parameters was performed. 54 subjects were divided into three groups in virtue of pancreatic region, three groups according to age, two groups according to gender, which pharmacokinetic parameters among and between different groups were calculated and compared. Results There was excellent agreement and low variability of intra- and inter-observer to pancreatic DCE-MRI pharmacokinetic parameters. The intra- and inter-observer ICCs of Ktrans, kep, ve, vp were 0.971, 0.952, 0.959, 0.944 and 0.947, 0.911, 0.978, 0.917, respectively. The intra- and inter-observer CoVs of Ktrans, kep, ve, vp were 9.98%, 5.99%, 6.47%, 4.76% and 10.15%, 5.22%, 6.28%, 5.40%, respectively. There were no significant differences of Ktrans, kep among different pancreatic regions, among different age groups, between male and female groups (P all > 0.10). Only, pancreatic ve of old group was higher than that of young and middle-aged groups (P = 0.042, 0.001), and vp of pancreatic head was higher than that of pancreatic body and tail (P = 0.014, 0.043). Conclusions DCE-MRI with ETL model is reliable and reproducible for quantitative assessment of pancreatic pharmacokinetic parameters. ve varies with age and vp varies with pancreatic region, which can provide guidance for the selection of normal reference in the pharmacokinetics study of pancreatic diseases.

Generating an interpretable and compact representation of 3D shapes from point clouds is an important and challenging problem. This paper presents CSG-Stump Net, an unsupervised end-to-end network for learning shapes from point clouds and discovering the underlying constituent modeling primitives and operations as well. At the core is a three-level structure called {\em CSG-Stump}, consisting of a complement layer at the bottom, an intersection layer in the middle, and a union layer at the top. CSG-Stump is proven to be equivalent to CSG in terms of representation, therefore inheriting the interpretable, compact and editable nature of CSG while freeing from CSG's complex tree structures. Particularly, the CSG-Stump has a simple and regular structure, allowing neural networks to give outputs of a constant dimensionality, which makes itself deep-learning friendly. Due to these characteristics of CSG-Stump, CSG-Stump Net achieves superior results compared to previous CSG-based methods and generates much more appealing shapes, as confirmed by extensive experiments. Project page: https://kimren227.github.io/projects/CSGStump/, Accepted to ICCV 2021

In the field of robotic hand design, soft body and anthropomorphic design are two trends with a promising future. Designing soft body anthropomorphic robotic hands with human-like grasping ability, but with a simple and reliable structure, is a challenge that still has not been not fully solved. In this paper, we present an anatomically correct robotic hand 3D model that aims to realize the human hand's functionality using a single type of 3D-printable material. Our robotic hand 3D model is combined with bones, ligaments, tendons, pulley systems, and tissue. We also describe the fabrication method to rapidly produce our robotic hand in 3D printing, wherein all parts are made by elastic 50 A (shore durometer) resin. In the experimental section, we show that our robotic hand has a similar motion range to a human hand with substantial grasping strength and compare it with the latest other designs of anthropomorphic robotic hands. Our new design greatly reduces the fabrication cost and assembly time. Compared with other robotic hand designs, we think our robotic hand may induce a new approach to the design and production of robotic hands as well as other related mechanical structures. National Research Foundation (NRF) Published version This research is supported by the National Research Foundation, Singapore, under its International Research Centres in Singapore Funding Initiative.

The circadian clock participates in maintaining homeostasis in peripheral tissues, including intervertebral discs (IVDs). Abnormal mechanical loading is a known risk factor for intervertebral disc degeneration (IDD). Based on the rhythmic daily loading pattern of rest and activity, we hypothesized that abnormal mechanical loading could dampen the IVD clock, contributing to IDD. Here, we investigated the effects of abnormal loading on the IVD clock and aimed to inhibit compression-induced IDD by targeting the core clock molecule brain and muscle Arnt-like protein-1 (BMAL1). In this study, we showed that BMAL1 KO mice exhibit radiographic features similar to those of human IDD and that BMAL1 expression was negatively correlated with IDD severity by systematic analysis based on 149 human IVD samples. The intrinsic circadian clock in the IVD was dampened by excessive loading, and BMAL1 overexpression by lentivirus attenuated compression-induced IDD. Inhibition of the RhoA/ROCK pathway by Y-27632 or melatonin attenuated the compression-induced decrease in BMAL1 expression. Finally, the two drugs partially restored BMAL1 expression and alleviated IDD in a diurnal compression model. Our results first show that excessive loading dampens the circadian clock of nucleus pulposus tissues via the RhoA/ROCK pathway, the inhibition of which potentially protects against compression-induced IDD by preserving BMAL1 expression. These findings underline the importance of the circadian clock for IVD homeostasis and provide a potentially effective therapeutic strategy for IDD.

Current research of designing prosthetic robotic hands mainly focuses on improving their functionality by devising new mechanical structures and actuation systems. Most of existing work relies on a single structure/system (e.g., bone-only or tissue-only) and ignores the fact that the human hand is composed of multiple functional structures (e.g., skin, bones, muscles, and tendons). This may increase the difficulty of the design process and lower the flexibility of the fabricated hand. To tackle this problem, this paper proposes a three-dimensional (3D) printable multi-layer design that models the hand with the layers of skin, tissues, and bones. The proposed design first obtains the 3D surface model of a target hand via 3D scanning, and then generates the 3D bone models from the surface model based on a fast template matching method. To overcome the disadvantage of the rigid bone layer in deformation, the tissue layer is introduced and represented by a concentric tube-based structure, of which the deformability can be explicitly controlled by a parameter. The experimental results show that the proposed design outperforms previous designs remarkably. With the proposed design, prosthetic robotic hands can be produced quickly with low cost and be customizable and deformable.

With the powerfulness of convolution neural networks (CNN), CNN based face reconstruction has recently shown promising performance in reconstructing detailed face shape from 2D face images. The success of CNN-based methods relies on a large number of labeled data. The state-of-the-art synthesizes such data using a coarse morphable face model, which however has difficulty to generate detailed photo-realistic images of faces (with wrinkles). This paper presents a novel face data generation method. Specifically, we render a large number of photo-realistic face images with different attributes based on inverse rendering. Furthermore, we construct a fine-detailed face image dataset by transferring different scales of details from one image to another. We also construct a large number of video-type adjacent frame pairs by simulating the distribution of real video data. With these nicely constructed datasets, we propose a coarse-to-fine learning framework consisting of three convolutional networks. The networks are trained for real-time detailed 3D face reconstruction from monocular video as well as from a single image. Extensive experimental results demonstrate that our framework can produce high-quality reconstruction but with much less computation time compared to the state-of-the-art. Moreover, our method is robust to pose, expression and lighting due to the diversity of data., Comment: Accepted by IEEE Transactions on Pattern Analysis and Machine Intelligence, 2018

Sleep deprivation (SD) has become very common in contemporary society, where people work around the clock. SD-induced cognitive deficits show large inter-individual differences and are trait-like with known neural correlates. However, few studies have used neuroimaging to predict vulnerability to SD. Here, resting state functional magnetic resonance imaging (fMRI) data and psychomotor vigilance task (PVT) data were collected from 60 healthy subjects after resting wakefulness and after one night of SD. The number of PVT lapses was then used to classify participants on the basis of whether they were vulnerable or resilient to SD. We explored the viability of graph-theory-based degree centrality to accurately classify vulnerability to SD. Compared with during resting wakefulness, widespread changes in degree centrality (DC) were found after SD, indicating significant reorganization of sleep homeostasis with respect to activity in resting state brain network architecture. Support vector machine (SVM) analysis using leave-one-out cross-validation achieved a correct classification rate of 84.75% [sensitivity 82.76%, specificity 86.67%, and area under the receiver operating characteristic curve (AUC) 0.94] for differentiating vulnerable subjects from resilient subjects. Brain areas that contributed most to the classification model were mainly located within the sensorimotor network, default mode network, and thalamus. Furthermore, we found a significantly negative correlation between changes in PVT lapses and DC in the thalamus after SD. These findings suggest that resting-state network measures combined with a machine learning algorithm could have broad potential applications in screening vulnerability to SD.

Most current anthropomorphic robotic hands can realize part of the human hand functions, particularly for object grasping. However, due to the complexity of the human hand, few current designs target at daily object manipulations, even for simple actions like rotating a pen. To tackle this problem, we introduce a gesture based framework, which adopts the widely-used 33 grasping gestures of Feix as the bases for hand design and implementation of manipulation. In the proposed framework, we first measure the motion ranges of human fingers for each gesture, and based on the results, we propose a simple yet dexterous robotic hand design with 13 degrees of actuation. Furthermore, we adopt a frame interpolation based method, in which we consider the base gestures as the key frames to represent a manipulation task, and use the simple linear interpolation strategy to accomplish the manipulation. To demonstrate the effectiveness of our framework, we define a three-level benchmark, which includes not only 62 test gestures from previous research, but also multiple complex and continuous actions. Experimental results on this benchmark validate the dexterity of the proposed design and our video is available in \url{https://drive.google.com/file/d/1wPtkd2P0zolYSBW7_3tVMUHrZEeXLXgD/view?usp=sharing}., Comment: 8 pages, 9 figures, 3 tables, accepted by IEEE Robotics and Automation Letters

We present a neural-based model for relighting a half-body portrait image by simply referring to another portrait image with the desired lighting condition. Rather than following classical inverse rendering methodology that involves estimating normals, albedo and environment maps, we implicitly encode the subject and lighting in a latent space, and use these latent codes to generate relighted images by neural rendering. A key technical innovation is the use of a novel overcomplete lighting representation, which facilitates lighting interpolation in the latent space, as well as helping regularize the self-organization of the lighting latent space during training. In addition, we propose a novel multiplicative neural render that more effectively combines the subject and lighting latent codes for rendering. We also created a large-scale photorealistic rendered relighting dataset for training, which allows our model to generalize well to real images. Extensive experiments demonstrate that our system not only outperforms existing methods for referral-based portrait relighting, but also has the capability generate sequences of relighted images via lighting rotations.

Motivated by how humans perceive geometry and color to recognize objects, we propose a Boundary Embedded Attentional Convolution (BEACon) network for point cloud instance segmentation. At the core of BEACon, we introduce the attentional weight in the convolution layer to adjust the neighboring features, with the weight being adapted to the relationship between geometry and color changes. As a result, BEACon makes use of both geometry and color information, takes instance boundary as an important feature, and thus learns a more discriminative feature representation in the neighborhood. Experimental results show that BEACon outperforms the state-of-the-art by a large margin. Ablation studies are also provided to prove the large benefit of incorporating both geometry and color into attention weight for instance segmentation. National Research Foundation (NRF) Accepted version This research is supported by the National Research Foundation, Singapore under its International Research Centres in Singapore Funding Initiative. Any opinions, findings and conclusions or recommendations expressed in this material are those of the author(s) and do not reflect the views of National Research Foundation, Singapore.

Corbel is a common category of decorative architectural geometry that has clear structure and aesthetic design. This paper presents a method for automatically generating a group of new corbel models from one selected by the user in the dataset. The method consists of offline learning and online generation. The offline learning trains two VAE models (2D CurveVAE and 3D VoxelVAE) for learning the feature representation of corbel parts. The online generation includes a generation algorithm by evolution that evolves to product new generation of models by crossing over and mutating features, and a feature-driven deformation that synthesizes 3D mesh representation of corbel models. By integrating these technical components, we develop a creative corbel modeling tool capable of generating new corbel models that are both “more of the same” and “surprising”, which is demonstrated by experiments. Ministry of Education (MOE) Accepted version This work is supported by the Ministry of Education, Singapore, under its MoE Tier-2 Grant (2017-T2-1-076).

Human bodies exhibit various shapes for different identities or poses, but the body shape has certain similarities in structure and thus can be embedded in a low-dimensional space. This paper presents an autoencoder-like network architecture to learn disentangled shape and pose embedding specifically for the 3D human body. This is inspired by recent progress of deformation-based latent representation learning. To improve the reconstruction accuracy, we propose a hierarchical reconstruction pipeline for the disentangling process and construct a large dataset of human body models with consistent connectivity for the learning of the neural network. Our learned embedding can not only achieve superior reconstruction accuracy but also provide great flexibility in 3D human body generation via interpolation, bilinear interpolation, and latent space sampling. The results from extensive experiments demonstrate the powerfulness of our learned 3D human body embedding in various applications., Comment: This manuscript is accepted for publication in the IEEE Transactions on Visualization and Computer Graphics Journal (IEEE TVCG). The Code is available at https://github.com/Juyong/DHNN_BodyRepresentation

It is known that both the number and the structure of somatic chromosomes can vary in early generation hexaploid wheats. The phenomenon is generally assumed to arise as a result of the meiotic instability characteristic of freshly created allopolyploids. Here, an analysis of the somatic karyotype of a set of 33 early generation synthetic hexaploid wheats has revealed that variation, taking the form of sub-chromosomal fragments and inter-chromosomal translocations, can also arise in somatic tissue. A possible explanation for the observations was that karyotypic instability in early generation hexaploid wheat probably occurs not just during sporogenesis, but also in somatic tissue. However, other factors such as the use of nitrous oxide during the experiments could also cause the chromosome variations, and additional experimentation would be required to determine the most likely.

In this paper, a model-free adaptive direct torque control (MFADTC) method for the speed regulation of asynchronous motors is proposed to solve the problems of modeling difficulties and poor anti-disturbance ability of the asynchronous motor. The designed model-free adaptive direct torque control (MFADTC) method depends merely on the input and the output data of the asynchronous motor. Numerical simulations are provided to show that this method has significantly improved the system&rsquo, s anti-disturbance ability.

Allopolyploidy increases its evolutionary potential by fixing heterosis and the advantage of gene redundancy. Allelic combinations generated from genetic recombination potentially provide many variations to the selection pools for evolution. May there be any relationship between allopolyploidization and genetic recombination? To study the impact of polyploidy on genetic recombination, we selected wheat as a model and simulated its evolution pathway of allopolyploidy by developing synthetic hexaploid wheat. The change of homologous chromosome recombination were investigated on their diploid DD and tetraploid AABB genomes after their allohexaploidization, respectively. The genetic recombination of the ancestral diploid genome ofAegilops tauschiiwas enhanced significantly more than 2 folds after their hexaploidization. Hexaploidization enhancing genetic recombination of the ancestral diploid D genome was firstly reported to be a new way to increase evolutionary potential of wheat, which is beneficial for wheat to conquer their narrow origination of D genome, quickly spread and make it a major crop of the world. Finally, re-synthetizing hexaploid wheat using diverseAe. tauschiispecies with tetraploid wheat can be considered as a pleiotropic strategy to speed adaptive evolution of bread wheat in breeding processes by increasing both gene allele types and genetic recombination variations.

Objectives To evaluate parameters of diffusion-weighted imaging (DWI) and dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) as early imaging indicators of tumor histologic response to pre-operative neoadjuvant chemotherapy and as probable prognostic factors for event-free survival (EFS) and overall survival in osteosarcoma (OS) in both tumoral and peritumoral areas. Methods Thirty-four OS patients who received three courses of neoadjuvant chemotherapy followed by surgery during 2014–2018 were enrolled in this study. All patients underwent baseline and post-chemotherapy DWI and DCE-MRI. Lesion region was defined as the tumoral area and peritumoral area. Parameters of apparent diffusion coefficient, capacity transfer constant (Ktrans), elimination rate constant, extravascular extracellular space volume ratio (Ve), and initial area under the curve as well as corresponding differences between pre- and post-chemotherapy in lesion regions were evaluated. Receiver operating characteristic analysis was used to evaluate the diagnostic performance of these parameters. The associations of all parameters with tumor histologic response, EFS, and overall survival were also calculated. Results In the tumor area, moderate evidence was found that post-Ktrans was lower in responders as compared with that in poor responders (p = 0.04, false discovery rate [FDR] corrected), and ΔKtrans exhibited significant between-groups differences (p = 0.04, Bonferroni corrected; or p = 0.006, FDR corrected). Weak evidence for the between-groups difference was found in the Ve in the peritumoral area (p = 0.025 before treatment and p = 0.021 after treatment, uncorrected). Furthermore, lower post-Ktrans in the tumoral area and lower pre-Ve in the peritumoral area were significant prognostic indicators for longer EFS (p = 0.002, p = 0.026) and overall survival (p = 0.003, p = 0.023). Conclusions In OS, DWI and DCE-MRI parameters in both tumoral and peritumoral areas can reflect the chemotherapy response and prognosticate EFS and overall survival. Key Points • Peritumoral MRI parameters can reflect the chemotherapy response in OS patients. • Peritumoral MRI parameters can predict EFS and overall survival in OS patients. • MRI parameters may be predictive factors for evaluating chemotherapy efficacy and EFS.

Deforming a tetrahedral mesh to conform to geometry modifications is a useful process in applications. This paper presents a method for tetrahedral mesh deformation driven by displacement of partial vertices of the mesh. The basic techniques behind the method are radial basis function (RBF)-based interpolation and adaptive mesh refinement. The method is realized by a warping process that transforms the mesh via iterative RBF-based interpolation to avoid the inversion of tetrahedra. Adaptive refinement by locally bisecting potentially inverted tetrahedra is also introduced to assure sufficiently large warping stepsizes. The refinement is performed on both the input mesh and the warped meshes concurrently to maintain the consistency of the topology. As a result, the method can effectively produce an inversion-free and topology compatible deformation mesh that satisfies hard positional constraints. Experimental results show the effectiveness of the method. Ministry of Education (MOE) This work was conducted in collaboration with HP Inc. and partially supported by the Singapore Government through the Industry Alignment Fund Industry Collaboration Projects Grant. It was also partially supported by the Ministry of Education, Singapore, under its MOE Tier-2 Grant (2017-T2-1-076).