A Large Model-Derived Algorithm for Complex Organoids with Internal Morphogenesis and Digital Marker Derivation.

Automated segmentation and evaluation algorithms have been demonstrated to enhance the simplicity and translational utility of organoid technology. However, there is a pressing need for the development of complex organoids that possess epithelium environmental elements, dense regional cell aggregation, and intraorganoid morphologies. Nevertheless, there has been limited progress, including both the construction of data sets and the development of algorithms, in the use of user-friendly microscopy to address such complex organoids. In this study, a data set of bright-field and living cell fluorescence images in paired forms and with temporal variance was constructed using droplet-engineered lung organoids. Additionally, a large model-based algorithm was developed. Both the organoid contours and intraorganoid morphologies were included in the data set, and their physical parameters were included and screened to form multiplex digital markers for organoid evaluation. The algorithm has been demonstrated to outperform existing methods and is therefore suitable for the evaluation of complex organoids. It is expected that the algorithm will facilitate the successful demonstration of AI in organoid evaluation and decision-making regarding their status.

Dispersion mismatch correction for evident chromatic anomaly in low coherence interferometry.

The applications of ultrafast optics to biomedical microscopy have expanded rapidly in recent years, including interferometric techniques like optical coherence tomography and microscopy (OCT/OCM). The advances of ultra-high resolution OCT and the inclusion of OCT/OCM in multimodal systems combined with multiphoton microscopy have marked a transition from using pseudo-continuous broadband sources, such as superluminescent diodes, to ultrafast supercontinuum optical sources. We report anomalies in the dispersion profiles of low-coherence ultrafast pulses through long and non-identical arms of a Michelson interferometer that are well beyond group delay or third-order dispersions. This chromatic anomaly worsens the observed axial resolution and causes fringe artifacts in the reconstructed tomograms in OCT/OCM using traditional algorithms. We present DISpersion COmpensation Techniques for Evident Chromatic Anomalies (DISCOTECA) as a universal solution to address the problem of chromatic dispersion mismatch in interferometry, especially with ultrafast sources. First, we demonstrate the origin of these artifacts through the self-phase modulation of ultrafast pulses due to focusing elements in the beam path. Next, we present three solution paradigms for DISCOTECA: optical, optoelectronic, and computational, along with quantitative comparisons to traditional methods to highlight the improvements to the dynamic range and axial profile. We explain the piecewise reconstruction of the phase mismatch between the arms of the spectral-domain interferometer using a modified short-term Fourier transform algorithm inspired by spectroscopic OCT. Finally, we present a decision-making guide for evaluating the utility of DISCOTECA in interferometry and for the artifact-free reconstruction of OCT images using an ultrafast supercontinuum source for biomedical applications.

USP21-mediated G3BP1 stabilization accelerates proliferation and metastasis of esophageal squamous cell carcinoma via activating Wnt/ß-Catenin signaling.

Lacking effective therapeutic targets heavily restricts the improvement of clinical prognosis for patients diagnosed with esophageal squamous cell carcinoma (ESCC). Ubiquitin Specific Peptidase 21 (USP21) is dysregulated in plenty of human cancers, however, its potential function and relevant molecular mechanisms in ESCC malignant progression as well as its value in clinical translation remain largely unknown. Here, in vitro and in vivo experiments revealed that aberrant upregulation of USP21 accelerated the proliferation and metastasis of ESCC in a deubiquitinase-dependent manner. Mechanistically, we found that USP21 binds to, deubiquitinates, and stabilizes the G3BP Stress Granule Assembly Factor 1 (G3BP1) protein, which is required for USP21-mediated ESCC progression. Further molecular studies demonstrated that the USP21/G3BP1 axis played a tumor-promoting role in ESCC progression by activating the Wnt/ß-Catenin signaling pathway. Additionally, disulfiram (DSF), an inhibitor against USP21 deubiquitylation activity, markedly abolished the USP21-mediated stability of G3BP1 protein and significantly displayed an anti-tumor effect on USP21-driving ESCC progression. Finally, the regulatory axis of USP21/G3BP1 was demonstrated to be aberrantly activated in ESCC tumor tissues and closely associated with advanced clinical stages and unfavorable prognoses, which provides a promising therapeutic strategy targeting USP21/G3BP1 axis for ESCC patients.

Programmable hyperspectral coherent anti-Stokes Raman scattering microscopy.

Hyperspectral coherent Raman scattering microscopy provides a significant improvement in acquisition time compared to spontaneous Raman scattering yet still suffers from the time required to sweep through individual wavenumbers. To address this, we present the use of a pulse shaper with a 2D spatial light modulator for phase- and amplitude-based shaping of the Stokes beam to create programmable spectrally tailored excitation envelopes. This enables collection of useful spectral information in a more rapid and efficient manner.

New insights into the influences of firework combustion on molecular composition and formation of sulfur- and halogen-containing organic compounds.

Firework (FW) events occur during various festivals worldwide and substantially negatively influence both air quality and human health. However, the effects of FWs on the chemical properties and formation of organic aerosols are far from clear. In this study, fine particulate matter (PM2.5) samples were collected in a suburban area in Qingdao, China during the Chinese Spring Festival. The concentrations of chemical species (especially carbonaceous components) in PM2.5 were measured using a combination of several state-of-the-art techniques. Our results showed that mass concentrations of water-soluble sulfate, potassium and chloride ions, and organic carbon drastically increased and became the predominant components in PM2.5 during FW events. Correspondingly, both the number and fractional contributions of sulfur (S)-containing subgroups (e.g., CHOS and CHONS compounds) and some chlorine (Cl)-containing organic (e.g., CHOSCl and CHONSCl) compounds identified using ultrahigh-resolution Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS) increased. The S- and Cl-containing compounds unique to the FW display period were identified, and their chemical characterization, sources, and formation mechanisms were elucidated by combining FT-ICR MS and quantum chemical calculations. Our results suggest that FW emissions play notable roles in both primary and secondary organic aerosol formation, especially for CHOS- and Cl-containing organic compounds.

Occurrence and seasonal variations of organophosphate flame retardants in air and dust from college microenvironments at Qingdao, China: Implications for student's exposure and risk assessment.

Organophosphate flame retardants (OPFRs) are widely used as alternatives to brominated flame retardants in a variety of consumer products and their consumption has continuously increased in recent years. However, their concentrations and human exposures in indoor microenvironments, particularly in a university environment, have received limited attention. In this study, the concentrations and seasonal variations of 15 OPFRs were assessed in typical microenvironments of two universities, including dormitories, offices, public microenvironments (PMEs: classroom, dining hall, gymnasium and library), and laboratories on the northern coast of China. Analysis of the OPFRs in both air and dust samples indicated widespread distribution in college campuses. The average concentration of ∑15OPFRs in the winter (12,774.4 ng/g and 5.3 ng/m3 for dust and air, respectively) was higher than in the summer (2460.4 ng/g and 4.6 ng/m3 for dust and air, respectively). The dust and air samples collected from PMEs and laboratories exhibited higher concentrations of OPFRs, followed by offices and dormitories. An equilibrium was reached between dust and air in all collected microenvironments. The daily intakes of OPFRs were significantly lower than the reference dose. Dust ingestion was the primary intake pathway in the winter, while inhalation and dust ingestion were the main intake pathways in the summer. The non-carcinogenic hazard quotients fell within the range of 10-7-10-3 in both the summer and winter, which are below the theoretical risk threshold. For the carcinogenic risk, the LCR values ranged from 10-10 to 10-8, indicating no elevated carcinogenic risk due to TnBP, TCEP, and TDCP in indoor dust and air.

Temporally optimized and spectrally shaped hyperspectral coherent anti-Stokes Raman scattering microscopy.

Coherent anti-Stokes Raman scattering (CARS) microscopy offers label-free chemical contrasts based on molecular vibrations. Hyperspectral CARS (HS-CARS) microscopy enables comprehensive microscale chemical characterization of biological samples. Various HS-CARS methods have been developed with individual advantages and disadvantages. We present what we believe to be a new temporally optimized and spectrally shaped (TOSS) HS-CARS method to overcome the limitations of existing techniques by providing precise control of the spatial and temporal profiles of the excitation beams for efficient and accurate measurements. This method uniquely uses Fourier transform pulse shaping based on a two-dimensional spatial light modulator to control the phase and amplitude of the excitation beams. TOSS-HS-CARS achieves fast, stable, and flexible acquisition, minimizes photodamage, and is highly adaptable to a multimodal multiphoton imaging system.

Analog multiplexing of a laser clock and computational photon counting for fast fluorescence lifetime imaging microscopy.

The dynamic range and fluctuations of fluorescence intensities and lifetimes in biological samples are large, demanding fast, precise, and versatile techniques. Among the high-speed fluorescence lifetime imaging microscopy (FLIM) techniques, directly sampling the output of analog single-photon detectors at GHz rates combined with computational photon counting can handle a larger range of photon rates. Traditionally, the laser clock is not sampled explicitly in fast FLIM; rather the detection is synchronized to the laser clock so that the excitation pulse train can be inferred from the cumulative photon statistics of several pixels. This has two disadvantages for sparse or weakly fluorescent samples: inconsistencies in inferring the laser clock within a frame and inaccuracies in aligning the decay curves from different frames for averaging. The data throughput is also very inefficient in systems with repetition rates much larger than the fluorescence lifetime due to significant silent regions where no photons are expected. We present a method for registering the photon arrival times to the excitation using time-domain multiplexing for fast FLIM. The laser clock is multiplexed with photocurrents into the silent region. Our technique does not add to the existing data bottleneck, has the sub-nanosecond dead time of computational photon counting based fast FLIM, works with various detectors, lasers, and electronics, and eliminates the errors in lifetime estimation in photon-starved conditions. We demonstrate this concept on two multiphoton setups of different laser repetition rates for single and multichannel FLIM multiplexed into a single digitizer channel for real-time imaging of biological samples.

Stabilization of MOF (KAT8) by USP10 promotes esophageal squamous cell carcinoma proliferation and metastasis through epigenetic activation of ANXA2/Wnt signaling.

Dysregulation of MOF (also known as MYST1, KAT8), a highly conserved H4K16 acetyltransferase, plays important roles in human cancers. However, its expression and function in esophageal squamous cell carcinoma (ESCC) remain unknown. Here, we report that MOF is highly expressed in ESCC tumors and predicts a worse prognosis. Depletion of MOF in ESCC significantly impedes tumor growth and metastasis both in vitro and in vivo, whereas ectopic expression of MOF but not catalytically inactive mutant (MOF-E350Q) promotes ESCC progression, suggesting that MOF acetyltransferase activity is crucial for its oncogenic activity. Further analysis reveals that USP10, a deubiquitinase highly expressed in ESCC, binds to and deubiquitinates MOF at lysine 410, which protects it from proteosome-dependent protein degradation. MOF stabilization by USP10 promotes H4K16ac enrichment in the ANXA2 promoter to stimulate ANXA2 transcription in a JUN-dependent manner, which subsequently activates Wnt/ß-Catenin signaling to facilitate ESCC progression. Our findings highlight a novel USP10/MOF/ANXA2 axis as a promising therapeutic target for ESCC.

Influence of oil extraction on concentration distributions, migration, secondary formation and carcinogenic risk of NPAHs and OPAHs in air and soil in an oilfield development area in China.

Oil extraction leads to environmental pollution from the oilfields and dweller activities, however, knowledge of the concentration distributions, migration, secondary formation and toxicity of nitrated/oxygenated polycyclic aromatic hydrocarbons (N/OPAHs) in oilfield regions is limited. In this research, atmospheric and soil samples in 7 different location types in an important oil industrial base in China were gathered. The ΣNPAHs and ΣOPAHs in the air ranged from 0.05 to 2.47 ng/m3 and 0.14-22.72 ng/m3, respectively, and in soil ranged from 0.22 to 17.81 ng/g and 9.69-66.86 ng/g, respectively. Both NPAHs and OPAHs in the atmosphere exhibited higher concentrations during winter. The atmospheric NPAH concentrations decreased exponentially with distance from urban area especially in the summer, revealing the impact of vehicles on the air in the Yellow River Delta area. High NPAH and OPAH concentrations were found only in soil near oil extraction facilities, indicating that the impact of oil extraction is limited to the soil near the extraction facilities. The air-soil exchanges of N/OPAHs were assessed through fugacity fraction analysis, and NPAHs were in the equilibrium-deposition state and OPAHs were in the net-deposition state in the winter. Higher incremental lifetime cancer risk (ILCR) occurred at the urban, industrial, and oilfield sites in the atmospheric samples, and the soil samples had the largest ILCR values in the oilfield sites. However, ILCR values for both air and soil did not exceed the threshold of 10-6.

[Plectin Promotes the Migration of Hepatocellular Carcinoma Cells Through Inducing F-actin Polymerization]. / Plectin通过诱导F-actin聚合增强肝癌细胞的迁移能力.

Objective: To explore the relationship between the expression of plectin and the migration of hepatocellular carcinoma (HCC) cells and to elucidate the molecular mechanisms by which plectin expression affects the migration of HCC cells. Methods: First of all, Western blot was performed to determine the expression of plectin in normal hepatocytes and HCC cells. Secondly, a plectin-downregulated HCC cell strain was established and the control group (shNC group) and shPLEC group were set up. Each group was divided into a vehicle control group (shNC+DMSO group or shPLEC+DMSO group) and a F-actin cytoskeleton polymerization inducer Jasplakinolide group (shNC+Jasp group or shPLEC+Jasp group). Western blot was performed to determine the expression of plectin and epithelial-mesenchymal transition (EMT)-related proteins, including N-cadherin, vimentin, and E-cadherin. HCC cell migration was evaluated by Transwell assay. KEGG (Kyoto Encyclopedia of Genes and Genomes) was used to analyze the signaling pathways related to plectin gene. The polymerization of F-actin was analyzed by immunofluorescence assay. Results: Compared with the normal hepatocytes, HCC cells showed high expression of plectin. Compared with those in the shNC group, the expression of plectin in the shPLEC group was decreased (P<0.05), the migration ability of HCC cells was weakened (P<0.05), and the EMT process was inhibited (with the expression of N-cadherin and vimentin being decreased and the expression of E-cadherin being increased) (P<0.05). KEGG analysis showed that the regulation of cytoskeletal F-actin was most closely associated with plectin and cytoskeletal F-actin depolymerized in the shPLEC group. After treatment with Jasplakinolide, an inducer of F-actin cytoskeleton polymerization, the migration ability of HCC cells in the shPLEC+Jasp group was enhanced compared with that of shPLEC+DMSO group (P<0.05) and the EMT process was restored (with the expression of N-cadherin and vimentin being increased and the expression of E-cadherin being decreased) (P<0.05). In addition, the polymerization of cytoskeletal F-actin in HCC cells was also restored. Conclusion: Plectin is highly expressed in HCC cells. Plectin promotes the migration and the EMT of HCC cells through inducing F-actin polymerization.

Targeted inhibition of the HNF1A/SHH axis by triptolide overcomes paclitaxel resistance in non-small cell lung cancer.

Paclitaxel resistance is associated with a poor prognosis in non-small cell lung cancer (NSCLC) patients, and currently, there is no promising drug for paclitaxel resistance. In this study, we investigated the molecular mechanisms underlying the chemoresistance in human NSCLC-derived cell lines. We constructed paclitaxel-resistant NSCLC cell lines (A549/PR and H460/PR) by long-term exposure to paclitaxel. We found that triptolide, a diterpenoid epoxide isolated from the Chinese medicinal herb Tripterygium wilfordii Hook F, effectively enhanced the sensitivity of paclitaxel-resistant cells to paclitaxel by reducing ABCB1 expression in vivo and in vitro. Through high-throughput sequencing, we identified the SHH-initiated Hedgehog signaling pathway playing an important role in this process. We demonstrated that triptolide directly bound to HNF1A, one of the transcription factors of SHH, and inhibited HNF1A/SHH expression, ensuing in attenuation of Hedgehog signaling. In NSCLC tumor tissue microarrays and cancer network databases, we found a positive correlation between HNF1A and SHH expression. Our results illuminate a novel molecular mechanism through which triptolide targets and inhibits HNF1A, thereby impeding the activation of the Hedgehog signaling pathway and reducing the expression of ABCB1. This study suggests the potential clinical application of triptolide and provides promising prospects in targeting the HNF1A/SHH pathway as a therapeutic strategy for NSCLC patients with paclitaxel resistance. Schematic diagram showing that triptolide overcomes paclitaxel resistance by mediating inhibition of the HNF1A/SHH/ABCB1 axis.

Key toxic components and sources affecting oxidative potential of atmospheric particulate matter using interpretable machine learning: Insights from fog episodes.

Fog significantly affects the air quality and human health. To investigate the health effects and mechanisms of atmospheric fine particulate matter (PM2.5) during fog episodes, PM2.5 samples were collected from the coastal suburb of Qingdao during different seasons from 2021 to 2022, with the major chemical composition in PM2.5 analyzed. The oxidative potential (OP) of PM2.5 was determined using the dithiothreitol (DTT) method. A positive matrix factorization model was adopted for PM2.5. Interpretable machine learning (IML) was used to reveal and quantify the key components and sources affecting OP. PM2.5 exhibited higher oxidative toxicity during fog episodes. Water-soluble organic carbon (WSOC), NH4+, K+, and water-soluble Fe positively affected the enhancement of DTTV (volume-based DTT activity) during fog episodes. The IML analysis demonstrated that WSOC and K+ contributed significantly to DTTV, with values of 0.31 ± 0.34 and 0.27 ± 0.22 nmol min-1 m-3, respectively. Regarding the sources, coal combustion and biomass burning contributed significantly to DTTV (0.40 ± 0.38 and 0.39 ± 0.36 nmol min-1 m-3, respectively), indicating the significant influence of combustion-related sources on OP. This study provides new insights into the effects of PM2.5 compositions and sources on OP by applying IML models.

Robust global arrangement by coherent enhancement in Huygens-Fresnel traveling surface acoustic wave interference field.

The application of standing surface acoustic wave (SSAW) tweezers based on backpropagation superposition to achieve precise behavior manipulation of microscale cells and even nanoscale bacteria has been widely studied and industrialized. However, the structure requires multiple transducer components or full channel resonance. It is very challenging to design a simple structure for nano-control by complex acoustic field. In this study, a reflector-interdigital transducer (R-IDT) acoustofluidic device based on unilateral coherence enhancement is proposed to achieve SSAW definition features of periodic particle capture positions. The SAW device based on a unilateral transducer can not only generate leaky-SAW in water-filled microchannel, but also have a contribution of spherical waves in the vibration area of the substrate-liquid interface due to the Huygens-Fresnel diffractive principle. Both of them form a robust time-averaged spatial periodicity in the pressure potential gradient, accurately predicting the lateral spacing of these positions through acoustic patterning methods. Furthermore, a reflector based on Bragg-reflection is used to suppress backward transmitted SAW and enhance forward conducted SAW beams. By using a finite element model, R-IDT structure's amplitude enhances 60.78% compared to single IDT structure. The particle manipulation range of the diffractive acoustic field greatly improves, verified by experimental polystyrene microspheres. Besides, biocompatibility is conformed through red blood cells and Bacillus subtilis. We investigate the overall shift of periodic pressure field that can still occur when the phase changes. This work provides a simpler and low-cost solution for the application of acoustic tweezer in biological cell culture and filtering.

Distributions of PAHs, NPAHs, OPAHs, BrPAHs, and ClPAHs in air, bulk deposition, soil, and water in the Shandong Peninsula, China: Urban-rural gradient, interface exchange, and long-range transport.

A systematic study of the movement of PAHs (Polycyclic aromatic hydrocarbons) and their derivatives through air, soil, and water is key to understanding the exchange and transport mechanisms of these pollutants in the environment and for ultimately improving environmental quality. PAHs and their derivatives, such as nitrated PAHs (NPAHs), oxygenated PAHs (OPAHs), brominated PAHs (BrPAHs) and chlorinated PAHs (ClPAHs), were analyzed in air, bulk deposition, soil, and water samples collected from urban, rural, field, and background sites on the eastern coast of China. The goal was to investigate and discuss their spatiotemporal variations, exchange fluxes, and transport potential. The concentrations of PAHs and their derivatives in the air and bulk deposition displayed distinct seasonal patterns, with higher concentrations observed during the winter and spring and lower concentrations during the summer and autumn. NPAHs exhibited the opposite trend. Significant urban-rural gradients were observed for most of the PAHs and their derivatives. According to the air-soil fugacity calculations, 2-3 ring PAHs, BrPAHs, and ClPAHs were found to volatilize from the soil into the air, while 4-7 ring PAHs, OPAHs, and NPAHs deposited from the air into the soil. The air-water fugacity of the PAHs and their derivatives indicated that surface water was an important source for the ambient atmosphere in Qingdao. The characteristic travel distances (CTDs) and persistence (Pov) for atmospheric transport were much lower than that for the water samples, which may be due to the longer half-lives of PAHs and their derivatives in water. NPAHs and ClPAHs with long transport distances and strong persistence in water could lead to a significant impact on marine pollution.

Toward Intrinsic Adversarial Robustness Through Probabilistic Training.

Modern deep neural networks have made numerous breakthroughs in real-world applications, yet they remain vulnerable to some imperceptible adversarial perturbations. These tailored perturbations can severely disrupt the inference of current deep learning-based methods and may induce potential security hazards to artificial intelligence applications. So far, adversarial training methods have achieved excellent robustness against various adversarial attacks by involving adversarial examples during the training stage. However, existing methods primarily rely on optimizing injective adversarial examples correspondingly generated from natural examples, ignoring potential adversaries in the adversarial domain. This optimization bias can induce the overfitting of the suboptimal decision boundary, which heavily jeopardizes adversarial robustness. To address this issue, we propose Adversarial Probabilistic Training (APT) to bridge the distribution gap between the natural and adversarial examples via modeling the latent adversarial distribution. Instead of tedious and costly adversary sampling to form the probabilistic domain, we estimate the adversarial distribution parameters in the feature level for efficiency. Moreover, we decouple the distribution alignment based on the adversarial probability model and the original adversarial example. We then devise a novel reweighting mechanism for the distribution alignment by considering the adversarial strength and the domain uncertainty. Extensive experiments demonstrate the superiority of our adversarial probabilistic training method against various types of adversarial attacks in different datasets and scenarios.

Spatial Distribution, Sources, Air-Soil Exchange, and Health Risks of Parent PAHs and Derivative-Alkylated PAHs in Different Functional Areas of an Oilfield Area in the Yellow River Delta, North China.

The knowledge of the spatial distribution, sources, and air-soil exchange of polycyclic aromatic compounds (PACs) in an oilfield area is essential to the development of effective control practices of PAC pollution. In this study, 48 passive air samples and 24 soil samples were collected during 2018-2019 in seven functional areas (e.g., urban, oil field, suburban, industrial, agricultural, near pump units, and background) in the Yellow River Delta (YRD) where the Shengli Oilfield is located, and 18 parent polycyclic aromatic hydrocarbons (PAHs) and five alkylated-PAHs (APAHs) were analyzed from all the air and soil samples. The ΣPAHs in the air and soil ranged from 2.26 to 135.83 ng/m3 and 33.96 to 408.94 ng/g, while the ΣAPAHs in the atmosphere and soil ranged from 0.04 to 16.31 ng/m3 and 6.39 to 211.86 ng/g, respectively. There was a downward trend of atmospheric ΣPAH concentrations with increasing the distance from the urban area, while both ΣPAH and ΣAPAH concentrations in the soil decreased with distance from the oilfield area. PMF analyses show that for atmospheric PACs, coal/biomass combustion was the main contributor in urban, suburban, and agricultural areas, while crude production and processing source contributes more in the industrial and oilfield area. For PACs in soil, densely populated areas (industrial, urban, and suburban) are more affected by traffic sources, while oilfield and near-pump unit areas are under the impact of oil spills. The fugacity fraction (ff) results indicated that the soil generally emitted low-molecular-weight PAHs and APAHs and act as a sink for high-molecular-weight PAHs. The incremental lifetime cancer risk (ILCR) of Σ(PAH+APAH) in both the air and soil, were below the threshold (≤10-6) set by the US EPA.

Learning Shadow Removal From Unpaired Samples via Reciprocal Learning.

We focus on addressing the problem of shadow removal for an image, and attempt to make a weakly supervised learning model that does not depend on the pixelwise-paired training samples, but only uses the samples with image-level labels that indicate whether an image contains shadow or not. To this end, we propose a deep reciprocal learning model that interactively optimizes the shadow remover and the shadow detector to improve the overall capability of the model. On the one hand, shadow removal is modeled as an optimization problem with a latent variable of the detected shadow mask. On the other hand, a shadow detector can be trained using the prior from the shadow remover. A self-paced learning strategy is employed to avoid fitting to intermediate noisy annotation during the interactive optimization. Furthermore, a color-maintenance loss and a shadow-attention discriminator are both designed to facilitate model optimization. Extensive experiments on the pairwise ISTD dataset, SRD dataset, and unpaired USR dataset demonstrate the superiority of the proposed deep reciprocal model.

Pose Guided Person Image Generation via Dual-task Correlation and Affinity Learning.

Pose Guided Person Image Generation (PGPIG) is the task of transforming a person's image from the source pose to a target pose. Existing PGPIG methods often tend to learn an end-to-end transformation between the source image and the target image, but do not seriously consider two issues: 1) the PGPIG is an ill-posed problem, and 2) the texture mapping requires effective supervision. In order to alleviate these two challenges, we propose a novel method by incorporating Dual-task Pose Transformer Network and Texture Affinity learning mechanism (DPTN-TA). To assist the ill-posed source-to-target task learning, DPTN-TA introduces an auxiliary task, i.e., source-to-source task, by a Siamese structure and further explores the dual-task correlation. Specifically, the correlation is built by the proposed Pose Transformer Module (PTM), which can adaptively capture the fine-grained mapping between sources and targets and can promote the source texture transmission to enhance the details of the generated images. Moreover, we propose a novel texture affinity loss to better supervise the learning of texture mapping. In this way, the network is able to learn complex spatial transformations effectively. Extensive experiments show that our DPTN-TA can produce perceptually realistic person images under significant pose changes. Furthermore, our DPTN-TA is not limited to processing human bodies but can be flexibly extended to view synthesis of other objects, i.e., faces and chairs, outperforming the state-of-the-arts in terms of both LPIPS and FID. Our code is available at: https://github.com/PangzeCheung/Dual-task-Pose-Transformer-Network.