Integrated metabolomic and transcriptomic analyses provide insights into regulation mechanisms during bulbous stem development in the Chinese medicinal herb plant, Stephania kwangsiensis.

BACKGROUND: Stephania kwangsiensis Lo (Menispermaceae) is a well-known Chinese herbal medicine, and its bulbous stems are used medicinally. The storage stem of S. kwangsiensis originated from the hypocotyls. To date, there are no reports on the growth and development of S. kwangsiensis storage stems. RESULTS: The bulbous stem of S. kwangsiensis, the starch diameter was larger at the stable expanding stage (S3T) than at the unexpanded stage (S1T) or the rapidly expanding stage (S2T) at the three different time points. We used ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) and Illumina sequencing to identify key genes involved in bulbous stem development. A large number of differentially accumulated metabolites (DAMs) and differentially expressed genes (DEGs) were identified. Based on the differential expression profiles of the metabolites, alkaloids, lipids, and phenolic acids were the top three differentially expressed classes. Compared with S2T, significant changes in plant signal transduction and isoquinoline alkaloid biosynthesis pathways occurred at both the transcriptional and metabolic levels in S1T. In S2T compared with S3T, several metabolites involved in tyrosine metabolism were decreased. Temporal analysis of S1T to S3T indicated the downregulation of phenylpropanoid biosynthesis, including lignin biosynthesis. The annotation of key pathways showed an up-down trend for genes and metabolites involved in isoquinoline alkaloid biosynthesis, whereas phenylpropanoid biosynthesis was not completely consistent. CONCLUSIONS: Downregulation of the phenylpropanoid biosynthesis pathway may be the result of carbon flow into alkaloid synthesis and storage of lipids and starch during the development of S. kwangsiensis bulbous stems. A decrease in the number of metabolites involved in tyrosine metabolism may also lead to a decrease in the upstream substrates of phenylpropane biosynthesis. Downregulation of lignin synthesis during phenylpropanoid biosynthesis may loosen restrictions on bulbous stem expansion. This study provides the first comprehensive analysis of the metabolome and transcriptome profiles of S. kwangsiensis bulbous stems. These data provide guidance for the cultivation, breeding, and harvesting of S. kwangsiensis.

Preparation and Performance Study of n-Undecane Phase Change Cold Storage Material.

With the fast development of the cold chain transportation industry, the traditional refrigeration method results in significant energy consumption. To address the national call for energy saving and emission reduction, the search for a new type of energy storage material has already become a future development trend. According to the national standard GB/T28577 for the classification and basic requirements of cold chain logistics, the temperature in frozen logistics is typically below -18 °C. In this study, n-undecane with a phase change temperature of -26 °C is chosen as the core material of microcapsules. Poly(methyl methacrylate) is applied as the shell material, with n-undecane microcapsules being prepared through suspension polymerization for phase change cold storage materials (MEPCM). Using characterization techniques including SEM, DSC, FTIR, and laser particle size analysis, the effects of three types of emulsifiers (SMA, Tween-80, Tween-80/span-80 (70/30)), SMA emulsifier dosage, core-shell ratio, and emulsification rate on the thermal performance and micro-surface morphology of n-undecane/PMMA microcapsules were studied. The results indicate that when comparing SMA, Tween-80, and Tween-80/span-80 (70/30) as emulsifiers, the dodecane/PMMA microcapsules prepared with SMA emulsifier exhibit superior thermal performance and micro-surface morphology, possessing a complete core-shell structure. The optimal microstructure and the highest enthalpy of phase change, measuring 120.3 kJ/kg, are achieved when SMA is used as the emulsifier with a quantity of 7%, a core-to-wall ratio of 2.5:1, and an emulsification speed of 2000 rpm. After 200 hot and cold cycles, the enthalpy of phase change decreased by only 18.6 kJ/kg, indicating the MEPCM thermal performance and cycle life. In addition, these optimized microcapsules exhibit favorable microstructure, uniform particle size, and efficient energy storage, making them an excellent choice for the refrigeration and freezing sectors.

Exploring Point-BEV Fusion for 3D Point Cloud Object Tracking with Transformer.

With the prevalent use of LiDAR sensors in autonomous driving, 3D point cloud object tracking has received increasing attention. In a point cloud sequence, 3D object tracking aims to predict the location and orientation of an object in consecutive frames. Motivated by the success of transformers, we propose Point Tracking TRansformer (PTTR), which efficiently predicts high-quality 3D tracking results in a coarse-to-fine manner with the help of transformer operations. PTTR consists of three novel designs. 1) Instead of random sampling, we design Relation-Aware Sampling to preserve relevant points to the given template during subsampling. 2) We propose a Point Relation Transformer for effective feature aggregation and feature matching between the template and search region. 3) Based on the coarse tracking results, we employ a novel Prediction Refinement Module to obtain the final refined prediction through local feature pooling. In addition, motivated by the favorable properties of the Bird's-Eye View (BEV) of point clouds in capturing object motion, we further design a more advanced framework named PTTR++, which incorporates both the point-wise view and BEV representation to exploit their complementary effect in generating high-quality tracking results. PTTR++ substantially boosts the tracking performance on top of PTTR with low computational overhead. Extensive experiments over multiple datasets show that our proposed approaches achieve superior 3D tracking accuracy and efficiency. Code will be available at https://github.com/Jasonkks/PTTR.

One-Shot Action Recognition via Multi-Scale Spatial-Temporal Skeleton Matching.

One-shot skeleton action recognition, which aims to learn a skeleton action recognition model with a single training sample, has attracted increasing interest due to the challenge of collecting and annotating large-scale skeleton action data. However, most existing studies match skeleton sequences by comparing their feature vectors directly which neglects spatial structures and temporal orders of skeleton data. This paper presents a novel one-shot skeleton action recognition technique that handles skeleton action recognition via multi-scale spatial-temporal feature matching. We represent skeleton data at multiple spatial and temporal scales and achieve optimal feature matching from two perspectives. The first is multi-scale matching which captures the scale-wise semantic relevance of skeleton data at multiple spatial and temporal scales simultaneously. The second is cross-scale matching which handles different motion magnitudes and speeds by capturing sample-wise relevance across multiple scales. Extensive experiments over three large-scale datasets (NTU RGB+D, NTU RGB+D 120, and PKU-MMD) show that our method achieves superior one-shot skeleton action recognition, and outperforms SOTA consistently by large margins.

Vision-Language Models for Vision Tasks: A Survey.

Most visual recognition studies rely heavily on crowd-labelled data in deep neural networks (DNNs) training, and they usually train a DNN for each single visual recognition task, leading to a laborious and time-consuming visual recognition paradigm. To address the two challenges, Vision-Language Models (VLMs) have been intensively investigated recently, which learns rich vision-language correlation from web-scale image-text pairs that are almost infinitely available on the Internet and enables zero-shot predictions on various visual recognition tasks with a single VLM. This paper provides a systematic review of visual language models for various visual recognition tasks, including: (1) the background that introduces the development of visual recognition paradigms; (2) the foundations of VLM that summarize the widely-adopted network architectures, pre-training objectives, and downstream tasks; (3) the widely-adopted datasets in VLM pre-training and evaluations; (4) the review and categorization of existing VLM pre-training methods, VLM transfer learning methods, and VLM knowledge distillation methods; (5) the benchmarking, analysis and discussion of the reviewed methods; (6) several research challenges and potential research directions that could be pursued in the future VLM studies for visual recognition. A project associated with this survey has been created at https://github.com/jingyi0000/VLM_survey.

Experimental study on CO2 capture by MEA/n-butanol/H2O phase change absorbent.

Monoethanolamines (MEAs) are widely used for CO2 capture, but their regeneration energy consumption is very high. CO2 Phase change absorbents (CPCAs) can be converted into CO2-rich and CO2-lean phases after absorbing CO2, and the regeneration energy consumption can be reduced because only the CO2-rich phase is thermally desorbed. In this paper, a novel CPCA with the composition "MEA/n-butanol/H2O (MNBH)" is proposed. Compared with the reported MEA phase change absorbent, the MNBH absorbent has higher CO2 absorption capacity, smaller absorbent viscosity and CO2-rich phase volume. The MNBH absorbent has the highest CO2 absorption capacity of 2.5227 mol CO2 per mol amine at a mass ratio of 3 : 4 : 3. The CO2 desorption efficiency reaches 89.96% at 120 °C, and the CO2 regeneration energy consumption is 2.6 GJ tCO2-1, which is about 35% lower than that of the 30 wt% MEA absorbent. When the mass ratio of MNBH absorbent was 3 : 6 : 1, the CO2 recycling capacity was 4.1918 mol CO2 L-1, which is 76% higher than that of the conventional 30 wt% MEA absorbent. The phase change absorbent developed in this paper can reduce the desorbent volume by about 50% and has good absorption performance for CO2 in flue gas.

POCE: Pose-Controllable Expression Editing.

Facial expression editing has attracted increasing attention with the advance of deep neural networks in recent years. However, most existing methods suffer from compromised editing fidelity and limited usability as they either ignore pose variations (unrealistic editing) or require paired training data (not easy to collect) for pose controls. This paper presents POCE, an innovative pose-controllable expression editing network that can generate realistic facial expressions and head poses simultaneously with just unpaired training images. POCE achieves the more accessible and realistic pose-controllable expression editing by mapping face images into UV space, where facial expressions and head poses can be disentangled and edited separately. POCE has two novel designs. The first is self-supervised UV completion that allows to complete UV maps sampled under different head poses, which often suffer from self-occlusions and missing facial texture. The second is weakly-supervised UV editing that allows to generate new facial expressions with minimal modification of facial identity, where the synthesized expression could be controlled by either an expression label or directly transplanted from a reference UV map via feature transfer. Extensive experiments show that POCE can learn from unpaired face images effectively, and the learned model can generate realistic and high-fidelity facial expressions under various new poses.

Public perceived risks and benefits of carbon capture, utilization, and storage (CCUS): Scale development and validation.

As a critical technology to mitigate climate change, the large-scale implementation of carbon capture, utilization, and storage (CCUS) depends on both technological advancement and public acceptance, which is significantly influenced by the perceived risks and benefits. Existing studies, however, have yet to reach a consensus regarding the measurement of CCUS in these two aspects. To fill this gap, this paper develops and validates new scales based on four studies. Specifically, Study 1 generates the initial item pool based on a literature review and semi-structured interviews, and then invites experts to examine the content validity of these items; Study 2 identifies the dimensions and assesses the reliability and validity of the measures through an exploratory and confirmatory factor analysis; Study 3 conducts a one-way ANOVA to test known-group validity; and Study 4 employed structural equation modeling to evaluate the nomological validity. The results demonstrate the internal consistency, reliability, and construct validity of the new scales developed to measure CCUS. This study provides a valuable tool for investigating public perceptions of CCUS and can help policymakers develop future publicity strategies.

Remarkable N2 selectivity enhancement of NH3-SCO reactions over V0.5/Pt0.04/TiO2 catalyst through Cu-Er co-modification.

Pt-V bimetallic catalysts maybe promising substitutes to precious metal catalysts for selective catalytic oxidation (SCO) of NH3. But it remains a major challenge for Pt-V bimetallic catalysts to pursue high NH3 conversion rate and N2 selectivity simultaneously. In this work, both Cu and Er were adopted to modify V0.5/Pt0.04/TiO2 catalyst (denoted as V/PT), and the influences of Cu and Er doping amounts on NH3-SCO performance of V/PT catalysts were investigated systematically. The results indicated that the co-modification of Cu and Er imposed little influence on NH3 conversion efficiency, but significantly boosted N2 selectivity. Compared with other Cu-Er-modified V/PT catalysts, CEV/PT-4 catalyst exhibited outstanding NH3-SCO performance, which attained completely 100% NH3 conversion efficiency and > 90% N2 selectivity in the temperature range of 225-450 °C. It was significantly superior to the NH3-SCO performance of most previously reported catalysts. The characterization results indicated that the adequate doping amounts of Cu and Er resulted in an obvious enhancement on redox property and surface acidity of CEV/PT-4 catalyst. It also led to abundant Pt0 and surface chemisorbed oxygen species on catalyst surface, which facilitated the oxidation of NH3 to NOx and enhanced i-SCR reactions. In situ DRIFTS results showed that -NH2 species on the surface of CEV/PT-4 catalyst could actively react with nitrate species to generate N2 and H2O.

Cross-Domain Facial Expression Recognition via Contrastive Warm up and Complexity-Aware Self-Training.

Unsupervised cross-domain Facial Expression Recognition (FER) aims to transfer the knowledge from a labeled source domain to an unlabeled target domain. Existing methods strive to reduce the discrepancy between source and target domain, but cannot effectively explore the abundant semantic information of the target domain due to the absence of target labels. To this end, we propose a novel framework via Contrastive Warm up and Complexity-aware Self-Training (namely CWCST), which facilitates source knowledge transfer and target semantic learning jointly. Specifically, we formulate a contrastive warm up strategy via features, momentum features, and learnable category centers to concurrently learn discriminative representations and narrow the domain gap, which benefits domain adaptation by generating more accurate target pseudo labels. Moreover, to deal with the inevitable noise in pseudo labels, we develop complexity-aware self-training with a label selection module based on prediction entropy, which iteratively generates pseudo labels and adaptively chooses the reliable ones for training, ultimately yielding effective target semantics exploration. Furthermore, by jointly using the two mentioned components, our framework enables to effectively utilize the source knowledge and target semantic information by source-target co- training. In addition, our framework can be easily incorporated into other baselines with consistent performance improvements. Extensive experimental results on seven databases show the superior performance of the proposed method against various baselines.

Multimodal Image Synthesis and Editing: The Generative AI Era.

As information exists in various modalities in real world, effective interaction and fusion among multimodal information plays a key role for the creation and perception of multimodal data in computer vision and deep learning research. With superb power in modeling the interaction among multimodal information, multimodal image synthesis and editing has become a hot research topic in recent years. Instead of providing explicit guidance for network training, multimodal guidance offers intuitive and flexible means for image synthesis and editing. On the other hand, this field is also facing several challenges in alignment of multimodal features, synthesis of high-resolution images, faithful evaluation metrics, etc. In this survey, we comprehensively contextualize the advance of the recent multimodal image synthesis and editing and formulate taxonomies according to data modalities and model types. We start with an introduction to different guidance modalities in image synthesis and editing, and then describe multimodal image synthesis and editing approaches extensively according to their model types. After that, we describe benchmark datasets and evaluation metrics as well as corresponding experimental results. Finally, we provide insights about the current research challenges and possible directions for future research.

Study on the Effect of PVAc and Styrene on the Properties and Microstructure of MMA-Based Repair Material for Concrete.

Methyl methacrylate (MMA) material is considered to be a suitable material for repairing concrete crack, provided that its large volume shrinkage during polymerization is resolved. This study was dedicated to investigating the effect of low shrinkage additives polyvinyl acetate and styrene (PVAc + styrene) on properties of the repair material and further proposes the shrinkage reduction mechanism based on the data of FTIR spectra, DSC testing and SEM micrographs. The results showed that PVAc + styrene delayed the gel point during the polymerization, and the formation of two-phase structure and micropores compensated for the volume shrinkage of the material. When the proportion of PVAc + styrene was 12%, the volume shrinkage could be as low as 4.78%, and the shrinkage stress was reduced by 87.4%. PVAc + styrene improved the bending strength and fracture toughness of most ratios investigated in this study. When 12% PVAc + styrene was added, the 28 d flexural strength and fracture toughness of MMA-based repair material were 28.04 MPa and 92.18%, respectively. After long-term curing, the repair material added with 12% PVAc + styrene showed a good adhesion to the substrate, with a bonding strength greater than 4.1 MPa and the fracture surface appearing at the substrate after the bonding experiment. This work contributes to the obtaining of a MMA-based repair material with low shrinkage, while its viscosity and other properties also can meet the requirements for repairing microcracks.

Experimental study on carbon capture characteristics of marine engine exhaust gas by activated potassium carbonate absorbent.

Post-combustion carbon capture is a direct and effective way for onboard carbon capture. Therefore, it is important to develop onboard carbon capture absorbent that can both ensure a high absorption rate and reduce the energy consumption of the desorption process. In this paper, a K2CO3 solution was first established using Aspen Plus to simulate CO2 capture from the exhaust gases of a marine dual-fuel engine in diesel mode. The lean and rich CO2 loading results from the simulation were used to guide the selection and optimization of the activators used in the experiment. During the experiment, five amino acid salt activators including SarK, GlyK, ProK, LysK, and AlaK and four organic amine activators including MEA, PZ, AEEA, and TEPA were used. Experiments only considered the activation effect of CO2 loading between lean and rich conditions. The results showed that after adding a small amount of activator, the absorption rate of CO2 by the absorbent was greatly improved, and the activation effect of organic amine activators was stronger than that of amino acid salts. Among the amino acid salts, the SarK-K2CO3 composite solution showed the best performance in both absorption and desorption. Among the amino acid salts and the organic amino activators, SarK-K2CO3 showed the best performance in strengthening the CO2 desorption while PZ-K2CO3 enhanced the CO2 absorption process the most. In the study of the concentration ratio, it was found that when the mass concentration ratio was 1:1 for SarK:K2CO3 and PZ:K2CO3, the CO2 absorption and desorption processes improved well.

Unsupervised Point Cloud Representation Learning With Deep Neural Networks: A Survey.

Point cloud data have been widely explored due to its superior accuracy and robustness under various adverse situations. Meanwhile, deep neural networks (DNNs) have achieved very impressive success in various applications such as surveillance and autonomous driving. The convergence of point cloud and DNNs has led to many deep point cloud models, largely trained under the supervision of large-scale and densely-labelled point cloud data. Unsupervised point cloud representation learning, which aims to learn general and useful point cloud representations from unlabelled point cloud data, has recently attracted increasing attention due to the constraint in large-scale point cloud labelling. This paper provides a comprehensive review of unsupervised point cloud representation learning using DNNs. It first describes the motivation, general pipelines as well as terminologies of the recent studies. Relevant background including widely adopted point cloud datasets and DNN architectures is then briefly presented. This is followed by an extensive discussion of existing unsupervised point cloud representation learning methods according to their technical approaches. We also quantitatively benchmark and discuss the reviewed methods over multiple widely adopted point cloud datasets. Finally, we share our humble opinion about several challenges and problems that could be pursued in the future research in unsupervised point cloud representation learning.

Image-to-Character-to-Word Transformers for Accurate Scene Text Recognition.

Leveraging the advances of natural language processing, most recent scene text recognizers adopt an encoder-decoder architecture where text images are first converted to representative features and then a sequence of characters via 'sequential decoding'. However, scene text images suffer from rich noises of different sources such as complex background and geometric distortions which often confuse the decoder and lead to incorrect alignment of visual features at noisy decoding time steps. This paper presents I2C2W, a novel scene text recognition technique that is tolerant to geometric and photometric degradation by decomposing scene text recognition into two inter-connected tasks. The first task focuses on image-to-character (I2C) mapping which detects a set of character candidates from images based on different alignments of visual features in an non-sequential way. The second task tackles character-to-word (C2W) mapping which recognizes scene text by decoding words from the detected character candidates. The direct learning from character semantics (instead of noisy image features) corrects falsely detected character candidates effectively which improves the final text recognition accuracy greatly. Extensive experiments over nine public datasets show that the proposed I2C2W outperforms the state-of-the-art by large margins for challenging scene text datasets with various curvature and perspective distortions. It also achieves very competitive recognition performance over multiple normal scene text datasets.

Investigating Pose Representations and Motion Contexts Modeling for 3D Motion Prediction.

Predicting human motion from historical pose sequence is crucial for a machine to succeed in intelligent interactions with humans. One aspect that has been obviated so far, is the fact that how we represent the skeletal pose has a critical impact on the prediction results. Yet there is no effort that investigates across different pose representation schemes. We conduct an indepth study on various pose representations with a focus on their effects on the motion prediction task. Moreover, recent approaches build upon off-the-shelf RNN units for motion prediction. These approaches process input pose sequence sequentially and inherently have difficulties in capturing long-term dependencies. In this paper, we propose a novel RNN architecture termed AHMR (Attentive Hierarchical Motion Recurrent network) for motion prediction which simultaneously models local motion contexts and a global context. We further explore a geodesic loss and a forward kinematics loss for the motion prediction task, which have more geometric significance than the widely employed L2 loss. Interestingly, we applied our method to a range of articulate objects including human, fish, and mouse. Empirical results show that our approach outperforms the state-of-the-art methods in short-term prediction and achieves much enhanced long-term prediction proficiency, such as retaining natural human-like motions over 50 seconds predictions. Our codes are released.

Meta-DETR: Image-Level Few-Shot Detection With Inter-Class Correlation Exploitation.

Few-shot object detection has been extensively investigated by incorporating meta-learning into region-based detection frameworks. Despite its success, the said paradigm is still constrained by several factors, such as (i) low-quality region proposals for novel classes and (ii) negligence of the inter-class correlation among different classes. Such limitations hinder the generalization of base-class knowledge for the detection of novel-class objects. In this work, we design Meta-DETR, which (i) is the first image-level few-shot detector, and (ii) introduces a novel inter-class correlational meta-learning strategy to capture and leverage the correlation among different classes for robust and accurate few-shot object detection. Meta-DETR works entirely at image level without any region proposals, which circumvents the constraint of inaccurate proposals in prevalent few-shot detection frameworks. In addition, the introduced correlational meta-learning enables Meta-DETR to simultaneously attend to multiple support classes within a single feedforward, which allows to capture the inter-class correlation among different classes, thus significantly reducing the misclassification over similar classes and enhancing knowledge generalization to novel classes. Experiments over multiple few-shot object detection benchmarks show that the proposed Meta-DETR outperforms state-of-the-art methods by large margins. The implementation codes are publicly available at https://github.com/ZhangGongjie/Meta-DETR.

GMLight: Lighting Estimation via Geometric Distribution Approximation.

Inferring the scene illumination from a single image is an essential yet challenging task in computer vision and computer graphics. Existing works estimate lighting by regressing representative illumination parameters or generating illumination maps directly. However, these methods often suffer from poor accuracy and generalization. This paper presents Geometric Mover's Light (GMLight), a lighting estimation framework that employs a regression network and a generative projector for effective illumination estimation. We parameterize illumination scenes in terms of the geometric light distribution, light intensity, ambient term, and auxiliary depth, which can be estimated by a regression network. Inspired by the earth mover's distance, we design a novel geometric mover's loss to guide the accurate regression of light distribution parameters. With the estimated light parameters, the generative projector synthesizes panoramic illumination maps with realistic appearance and high-frequency details. Extensive experiments show that GMLight achieves accurate illumination estimation and superior fidelity in relighting for 3D object insertion. The codes are available at https://github.com/fnzhan/Illumination-Estimation.

Effect of Nanofiller on the Mechanical Properties of Carbon Fiber/Epoxy Composites under Different Aging Conditions.

In this study, carbon fiber-reinforced epoxy composites (CFRPs) containing multi-walled carbon nanotube (MWCNT) and halloysite nanoclay were fabricated. The effects of these nanofillers (MWCNT and nanoclay) on the tensile and flexural properties of the CFRPs under different aging conditions were studied. These aging conditions included water soaking, acid soaking, alkali soaking, and thermal shock cycling. The experimental results showed that, after accelerated aging, the mechanical performance of the CFRPs decreased. The performance degradation in the soaking environment depends on the immersion temperature and immersion medium. High-temperature accelerated the aging behavior of the CFRPs, resulting in low strength and modulus. The CFRPs were more vulnerable to acid soaking and alkali soaking than water soaking. The MWCNT and halloysite nanoclay are beneficial to improve the immersion aging resistance of the CFRPs, and the additions of nanofillers delayed the performance degradation under immersion aging conditions. However, nanofillers hardly improve the aging resistance of the CFRPs under thermal shock cycling condition. The fracture morphologies were observed by scanning electron microscopy (SEM) to reflect the failure modes of the CFRPs under various aging conditions. Differential scanning calorimeter (DSC) and fourier transform infrared (FTIR) spectroscopy tests were used to estimate the changes in the chemical structures and properties of epoxy resin and its composites under different conditions.

PoT-GAN: Pose Transform GAN for Person Image Synthesis.

Pose-based person image synthesis aims to generate a new image containing a person with a target pose conditioned on a source image containing a person with a specified pose. It is challenging as the target pose is arbitrary and often significantly differs from the specified source pose, which leads to large appearance discrepancy between the source and the target images. This paper presents the Pose Transform Generative Adversarial Network (PoT-GAN) for person image synthesis where the generator explicitly learns the transform between the two poses by manipulating the corresponding multi-scale feature maps. By incorporating the learned pose transform information into the multi-scale feature maps of the source image in a GAN architecture, our method reliably transfers the appearance of the person in the source image to the target pose with no need for any hard-coded spatial information depicting the change of pose. According to both qualitative and quantitative results, the proposed PoT-GAN demonstrates a state-of-the-art performance on three publicly available datasets for person image synthesis.