Shenzhen Institute of Information Technology

We propose a new approach for estimation of the positions of facial key points with three-level carefully designed convolutional networks. At each level, the outputs of multiple networks are fused for robust and accurate estimation. Thanks to the deep structures of convolutional networks, global high-level features are extracted over the whole face region at the initialization stage, which help to locate high accuracy key points. There are two folds of advantage for this. First, the texture context information over the entire face is utilized to locate each key point. Second, since the networks are trained to predict all the key points simultaneously, the geometric constraints among key points are implicitly encoded. The method therefore can avoid local minimum caused by ambiguity and data corruption in difficult image samples due to occlusions, large pose variations, and extreme lightings. The networks at the following two levels are trained to locally refine initial predictions and their inputs are limited to small regions around the initial predictions. Several network structures critical for accurate and robust facial point detection are investigated. Extensive experiments show that our approach outperforms state-of-the-art methods in both detection accuracy and reliability.

Existing blind image quality assessment (BIQA) methods are mostly opinion-aware. They learn regression models from training images with associated human subjective scores to predict the perceptual quality of test images. Such opinion-aware methods, however, require a large amount of training samples with associated human subjective scores and of a variety of distortion types. The BIQA models learned by opinion-aware methods often have weak generalization capability, hereby limiting their usability in practice. By comparison, opinion-unaware methods do not need human subjective scores for training, and thus have greater potential for good generalization capability. Unfortunately, thus far no opinion-unaware BIQA method has shown consistently better quality prediction accuracy than the opinion-aware methods. Here, we aim to develop an opinion-unaware BIQA method that can compete with, and perhaps outperform, the existing opinion-aware methods. By integrating the features of natural image statistics derived from multiple cues, we learn a multivariate Gaussian model of image patches from a collection of pristine natural images. Using the learned multivariate Gaussian model, a Bhattacharyya-like distance is used to measure the quality of each image patch, and then an overall quality score is obtained by average pooling. The proposed BIQA method does not need any distorted sample images nor subjective quality scores for training, yet extensive experiments demonstrate its superior quality-prediction performance to the state-of-the-art opinion-aware BIQA methods. The MATLAB source code of our algorithm is publicly available at www.comp.polyu.edu.hk/~cslzhang/IQA/ILNIQE/ILNIQE.htm.

Reports on bacteria detected in maternal fluids during pregnancy are typically associated with adverse consequences, and whether the female reproductive tract harbours distinct microbial communities beyond the vagina has been a matter of debate. Here we systematically sample the microbiota within the female reproductive tract in 110 women of reproductive age, and examine the nature of colonisation by 16S rRNA gene amplicon sequencing and cultivation. We find distinct microbial communities in cervical canal, uterus, fallopian tubes and peritoneal fluid, differing from that of the vagina. The results reflect a microbiota continuum along the female reproductive tract, indicative of a non-sterile environment. We also identify microbial taxa and potential functions that correlate with the menstrual cycle or are over-represented in subjects with adenomyosis or infertility due to endometriosis. The study provides insight into the nature of the vagino-uterine microbiome, and suggests that surveying the vaginal or cervical microbiota might be useful for detection of common diseases in the upper reproductive tract.Whether the female reproductive tract harbours distinct microbiomes beyond the vagina has been a matter of debate. Here, the authors show a subject-specific continuity in microbial communities at six sites along the female reproductive tract, indicative of a non-sterile environment.

Zn-air batteries are becoming the promising power sources for portable and wearable electronic devices and hybrid/electric vehicles because of their high specific energy density and the low cost for next-generation green and sustainable energy technologies. An air electrode integrated with an oxygen electrocatalyst is the most important component and inevitably determines the performance and cost of a Zn-air battery. This article presents exciting advances and challenges related to air electrodes and their relatives. After a brief introduction of the Zn-air battery, the architectures and oxygen electrocatalysts of air electrodes and relevant electrolytes are highlighted in primary and rechargeable types with different configurations, respectively. Moreover, the individual components and major issues of flexible Zn-air batteries are also highlighted, along with the strategies to enhance the battery performance. Finally, a perspective for design, preparation, and assembly of air electrodes is proposed for the future innovations of Zn-air batteries with high performance.

Automatic medical image segmentation has made great progress owing to the powerful deep representation learning. Inspired by the success of self-attention mechanism in Transformer, considerable efforts are devoted to designing the robust variants of encoder-decoder architecture with Transformer. However, the patch division used in the existing Transformer-based models usually ignores the pixel-level intrinsic structural features inside each patch. In this paper, we propose a novel deep medical image segmentation framework called Dual Swin Transformer U-Net (DS-TransUNet), which aims to incorporate the hierarchical Swin Transformer into both encoder and decoder of the standard U-shaped architecture. Our DS-TransUNet benefits from the self-attention computation in Swin Transformer and the designed dual-scale encoding, which can effectively model the non-local dependencies and multi-scale contexts for enhancing the semantic segmentation quality of varying medical images. Unlike many prior Transformer-based solutions, the proposed DS-TransUNet adopts a well-established dual-scale encoding mechanism that utilizes dual-scale encoders based on Swin Transformer to extract the coarse and fine-grained feature representations of different semantic scales. Meanwhile, a well-designed Transformer Interactive Fusion (TIF) module is proposed to effectively perform the multi-scale information fusion through the self-attention mechanism. Furthermore, we introduce the Swin Transformer block into decoder to further explore the long-range contextual information during the up-sampling process. Extensive experiments across four typical tasks for medical image segmentation demonstrate the effectiveness of DS-TransUNet, and our approach significantly outperforms the state-of-the-art methods.

Abstract The development of single molecule white light emitters is extremely challenging for pure phosphorescent metal-free system at room temperature. Here we report a single pure organic phosphor, namely 4-chlorobenzoyldibenzothiophene, emitting white room temperature phosphorescence with Commission Internationale de l’Éclair-age coordinates of (0.33, 0.35). Experimental and theoretical investigations reveal that the white light emission is emerged from dual phosphorescence, which emit from the first and second excited triplet states. We also demonstrate the validity of the strategy to achieve metal-free pure phosphorescent single molecule white light emitters by intrasystem mixing dual room temperature phosphorescence arising from the low- and high-lying triplet states.

Chaos is known as a natural candidate for cryptography applications owing to its properties such as unpredictability and initial state sensitivity. However, certain chaos-based cryptosystems have been proven to exhibit various security defects because their used chaotic maps do not have complex dynamical behaviors. To address this problem, this paper introduces a cosine-transform-based chaotic system (CTBCS). Using two chaotic maps as seed maps, the CTBCS can produce chaotic maps with complex dynamical behaviors. For illustration, we produce three chaotic maps using the CTBCS and analyze their chaos complexity. Using one of the generated chaotic maps, we further propose an image encryption scheme. The encryption scheme uses high-efficiency scrambling to separate adjacent pixels and employs random order substitution to spread a small change in the plain-image to all pixels of the cipher-image. The performance evaluation demonstrates that the chaotic maps generated by the CTBCS exhibit substantially more complicated chaotic behaviors than the existing ones. The simulation results indicate the reliability of the proposed image encryption scheme. Moreover, the security analysis demonstrates that the proposed image encryption scheme provides a higher level of security than several advanced image encryption schemes.

Xenotransplantation is a promising strategy to alleviate the shortage of organs for human transplantation. In addition to the concerns about pig-to-human immunological compatibility, the risk of cross-species transmission of porcine endogenous retroviruses (PERVs) has impeded the clinical application of this approach. We previously demonstrated the feasibility of inactivating PERV activity in an immortalized pig cell line. We now confirm that PERVs infect human cells, and we observe the horizontal transfer of PERVs among human cells. Using CRISPR-Cas9, we inactivated all of the PERVs in a porcine primary cell line and generated PERV-inactivated pigs via somatic cell nuclear transfer. Our study highlights the value of PERV inactivation to prevent cross-species viral transmission and demonstrates the successful production of PERV-inactivated animals to address the safety concern in clinical xenotransplantation.

With the avalanche of biological sequences generated in the post-genomic age, one of the most challenging problems in computational biology is how to effectively formulate the sequence of a biological sample (such as DNA, RNA or protein) with a discrete model or a vector that can effectively reflect its sequence pattern information or capture its key features concerned. Although several web servers and stand-alone tools were developed to address this problem, all these tools, however, can only handle one type of samples. Furthermore, the number of their built-in properties is limited, and hence it is often difficult for users to formulate the biological sequences according to their desired features or properties. In this article, with a much larger number of built-in properties, we are to propose a much more flexible web server called Pse-in-One (http://bioinformatics.hitsz.edu.cn/Pse-in-One/), which can, through its 28 different modes, generate nearly all the possible feature vectors for DNA, RNA and protein sequences. Particularly, it can also generate those feature vectors with the properties defined by users themselves. These feature vectors can be easily combined with machine-learning algorithms to develop computational predictors and analysis methods for various tasks in bioinformatics and system biology. It is anticipated that the Pse-in-One web server will become a very useful tool in computational proteomics, genomics, as well as biological sequence analysis. Moreover, to maximize users' convenience, its stand-alone version can also be downloaded from http://bioinformatics.hitsz.edu.cn/Pse-in-One/download/, and directly run on Windows, Linux, Unix and Mac OS.

Abstract Developing highly active and durable electrocatalysts for acidic oxygen evolution reaction remains a great challenge due to the sluggish kinetics of the four-electron transfer reaction and severe catalyst dissolution. Here we report an electrochemical lithium intercalation method to improve both the activity and stability of RuO 2 for acidic oxygen evolution reaction. The lithium intercalates into the lattice interstices of RuO 2 , donates electrons and distorts the local structure. Therefore, the Ru valence state is lowered with formation of stable Li-O-Ru local structure, and the Ru–O covalency is weakened, which suppresses the dissolution of Ru, resulting in greatly enhanced durability. Meanwhile, the inherent lattice strain results in the surface structural distortion of Li x RuO 2 and activates the dangling O atom near the Ru active site as a proton acceptor, which stabilizes the OOH* and dramatically enhances the activity. This work provides an effective strategy to develop highly efficient catalyst towards water splitting.

Rechargeable aqueous zinc–cobalt oxide batteries with high voltage, excellent rate capability and long-cycling life.

The outbreak of the COVID-19 pandemic was partially due to the challenge of identifying asymptomatic and presymptomatic carriers of the virus, and thus highlights a strong motivation for diagnostics with high sensitivity that can be rapidly deployed. On the other hand, several concerning SARS-CoV-2 variants, including Omicron, are required to be identified as soon as the samples are identified as 'positive'. Unfortunately, a traditional PCR test does not allow their specific identification. Herein, for the first time, we have developed MOPCS (Methodologies of Photonic CRISPR Sensing), which combines an optical sensing technology-surface plasmon resonance (SPR) with the 'gene scissors' clustered regularly interspaced short palindromic repeat (CRISPR) technique to achieve both high sensitivity and specificity when it comes to measurement of viral variants. MOPCS is a low-cost, CRISPR/Cas12a-system-empowered SPR gene-detecting platform that can analyze viral RNA, without the need for amplification, within 38 min from sample input to results output, and achieve a limit of detection of 15 fM. MOPCS achieves a highly sensitive analysis of SARS-CoV-2, and mutations appear in variants B.1.617.2 (Delta), B.1.1.529 (Omicron) and BA.1 (a subtype of Omicron). This platform was also used to analyze some recently collected patient samples from a local outbreak in China, identified by the Centers for Disease Control and Prevention. This innovative CRISPR-empowered SPR platform will further contribute to the fast, sensitive and accurate detection of target nucleic acid sequences with single-base mutations.

Abstract Background Rapid spread of SARS-CoV-2 in Wuhan prompted heightened surveillance in Shenzhen and elsewhere in China. The resulting data provide a rare opportunity to measure key metrics of disease course, transmission, and the impact of control. Methods The Shenzhen CDC identified 391 SARS-CoV-2 cases from January 14 to February 12, 2020 and 1286 close contacts. We compare cases identified through symptomatic surveillance and contact tracing, and estimate the time from symptom onset to confirmation, isolation, and hospitalization. We estimate metrics of disease transmission and analyze factors influencing transmission risk. Findings Cases were older than the general population (mean age 45) and balanced between males (187) and females (204). Ninety-one percent had mild or moderate clinical severity at initial assessment. Three have died, 225 have recovered (median time to recovery is 21 days). Cases were isolated on average 4.6 days after developing symptoms; contact tracing reduced this by 1.9 days. Household contacts and those travelling with a case where at higher risk of infection (ORs 6 and 7) than other close contacts. The household secondary attack rate was 15%, and children were as likely to be infected as adults. The observed reproductive number was 0.4, with a mean serial interval of 6.3 days. Interpretation Our data on cases as well as their infected and uninfected close contacts provide key insights into SARS-CoV-2 epidemiology. This work shows that heightened surveillance and isolation, particularly contact tracing, reduces the time cases are infectious in the community, thereby reducing R . Its overall impact, however, is uncertain and highly dependent on the number of asymptomatic cases. We further show that children are at similar risk of infection as the general population, though less likely to have severe symptoms; hence should be considered in analyses of transmission and control.

Long-read sequencing is promising for the comprehensive discovery of structural variations (SVs). However, it is still non-trivial to achieve high yields and performance simultaneously due to the complex SV signatures implied by noisy long reads. We propose cuteSV, a sensitive, fast, and scalable long-read-based SV detection approach. cuteSV uses tailored methods to collect the signatures of various types of SVs and employs a clustering-and-refinement method to implement sensitive SV detection. Benchmarks on simulated and real long-read sequencing datasets demonstrate that cuteSV has higher yields and scaling performance than state-of-the-art tools. cuteSV is available at https://github.com/tjiangHIT/cuteSV .

Abstract Aqueous zinc (Zn) batteries (AZBs) are widely considered as a promising candidate for next‐generation energy storage owing to their excellent safety features. However, the application of a Zn anode is hindered by severe dendrite formation and side reactions. Herein, an interfacial bridged organic–inorganic hybrid protection layer (Nafion‐Zn‐X) is developed by complexing inorganic Zn‐X zeolite nanoparticles with Nafion, which shifts ion transport from channel transport in Nafion to a hopping mechanism in the organic–inorganic interface. This unique organic–inorganic structure is found to effectively suppress dendrite growth and side reactions of the Zn anode. Consequently, the Zn@Nafion‐Zn‐X composite anode delivers high coulombic efficiency (ca. 97 %), deep Zn plating/stripping (10 mAh cm −2 ), and long cycle life (over 10 000 cycles). By tackling the intrinsic chemical/electrochemical issues, the proposed strategy provides a versatile remedy for the limited cycle life of the Zn anode.

The incorporation of hydroxyapatite (HAP) into poly-l-lactic acid (PLLA) matrix serving as bone scaffold is expected to exhibit bioactivity and osteoconductivity to those of the living bone. While too low degradation rate of HAP/PLLA scaffold hinders the activity because the embedded HAP in the PLLA matrix is difficult to contact and exchange ions with body fluid. In this study, biodegradable polymer poly (glycolic acid) (PGA) was blended into the HAP/PLLA scaffold fabricated by laser 3D printing to accelerate the degradation. The results indicated that the incorporation of PGA enhanced the degradation rate of scaffold as indicated by the weight loss increasing from 3.3% to 25.0% after immersion for 28 days, owing to the degradation of high hydrophilic PGA and the subsequent accelerated hydrolysis of PLLA chains. Moreover, a lot of pores produced by the degradation of the scaffold promoted the exposure of HAP from the matrix, which not only activated the deposition of bone like apatite on scaffold but also accelerated apatite growth. Cytocompatibility tests exhibited a good osteoblast adhesion, spreading and proliferation, suggesting the scaffold provided a suitable environment for cell cultivation. Furthermore, the scaffold displayed excellent bone defect repair capacity with the formation of abundant new bone tissue and blood vessel tissue, and both ends of defect region were bridged after 8 weeks of implantation.

Abstract Continuous and real‐time sensoring has received much attention in biomarker monitoring, toxicity assessment, and therapeutic agent tracking. However, its on‐site application is seriously limited by several stubborn defects including liability to fouling, signal drifting, short service life, poor repeatability, etc. Additionally, most current methods require extra sample pretreatment, delaying timely acquisition of testing results. To address these issues, MXene‐Ti 3 C 2 T x based screen‐printed electrode incorporated with a dialysis microfluidic chip is constructed for a direct and continuous multicomponent analysis of whole blood. Dual‐function of MXene is developed and allows for simultaneous quantification of different target compounds through one device. Importantly, ratiometric sensing tactic is easily implemented in the system, which greatly alleviates signal drifting. As a proof of concept, this novel sensor is applied in hemodialysis, and continuous assay of urea, uric acid, and creatinine levels in human blood is realized. This work paves a new path for 2D MXene in biomedical and sensing applications.

Ir-based binary and ternary alloys are effective catalysts for the electrochemical oxygen evolution reaction (OER) in acidic solutions. Nevertheless, decreasing the Ir content to less than 50 at% while maintaining or even enhancing the overall electrocatalytic activity and durability remains a grand challenge. Herein, by dealloying predesigned Al-based precursor alloys, it is possible to controllably incorporate Ir with another four metal elements into one single nanostructured phase with merely ≈20 at% Ir. The obtained nanoporous quinary alloys, i.e., nanoporous high-entropy alloys (np-HEAs) provide infinite possibilities for tuning alloy's electronic properties and maximizing catalytic activities owing to the endless element combinations. Particularly, a record-high OER activity is found for a quinary AlNiCoIrMo np-HEA. Forming HEAs also greatly enhances the structural and catalytic durability regardless of the alloy compositions. With the advantages of low Ir loading and high activity, these np-HEA catalysts are very promising and suitable for activity tailoring/maximization.

The exploitation of a high-efficient, low-cost, and stable non-noble-metal-based catalyst with oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) simultaneously, as air electrode material for a rechargeable zinc–air battery is significantly crucial. Meanwhile, the compressible flexibility of a battery is the prerequisite of wearable or/and portable electronics. Herein, we present a strategy via single-site dispersion of an Fe–Nx species on a two-dimensional (2D) highly graphitic porous nitrogen-doped carbon layer to implement superior catalytic activity toward ORR/OER (with a half-wave potential of 0.86 V for ORR and an overpotential of 390 mV at 10 mA·cm–2 for OER) in an alkaline medium. Furthermore, an elastic polyacrylamide hydrogel based electrolyte with the capability to retain great elasticity even under a highly corrosive alkaline environment is utilized to develop a solid-state compressible and rechargeable zinc–air battery. The creatively developed battery has a low charge–discharge voltage gap (0.78 V at 5 mA·cm–2) and large power density (118 mW·cm–2). It could be compressed up to 54% strain and bent up to 90° without charge/discharge performance and output power degradation. Our results reveal that single-site dispersion of catalytic active sites on a porous support for a bifunctional oxygen catalyst as cathode integrating a specially designed elastic electrolyte is a feasible strategy for fabricating efficient compressible and rechargeable zinc–air batteries, which could enlighten the design and development of other functional electronic devices.

In this paper, we propose a general framework for incomplete multiview clustering. The proposed method is the first work that exploits the graph learning and spectral clustering techniques to learn the common representation for incomplete multiview clustering. First, owing to the good performance of low-rank representation in discovering the intrinsic subspace structure of data, we adopt it to adaptively construct the graph of each view. Second, a spectral constraint is used to achieve the low-dimensional representation of each view based on the spectral clustering. Third, we further introduce a co-regularization term to learn the common representation of samples for all views, and then use the k -means to partition the data into their respective groups. An efficient iterative algorithm is provided to optimize the model. Experimental results conducted on seven incomplete multiview datasets show that the proposed method achieves the best performance in comparison with some state-of-the-art methods, which proves the effectiveness of the proposed method in incomplete multiview clustering.