Showing papers by "Jiawei Han published in 2022"

A fully automatic AI system for tooth and alveolar bone segmentation from cone-beam CT images

Abstract: Abstract Accurate delineation of individual teeth and alveolar bones from dental cone-beam CT (CBCT) images is an essential step in digital dentistry for precision dental healthcare. In this paper, we present an AI system for efficient, precise, and fully automatic segmentation of real-patient CBCT images. Our AI system is evaluated on the largest dataset so far, i.e., using a dataset of 4,215 patients (with 4,938 CBCT scans) from 15 different centers. This fully automatic AI system achieves a segmentation accuracy comparable to experienced radiologists (e.g., 0.5% improvement in terms of average Dice similarity coefficient), while significant improvement in efficiency (i.e., 500 times faster). In addition, it consistently obtains accurate results on the challenging cases with variable dental abnormalities, with the average Dice scores of 91.5% and 93.0% for tooth and alveolar bone segmentation. These results demonstrate its potential as a powerful system to boost clinical workflows of digital dentistry.

ESCM2: Entire Space Counterfactual Multi-Task Model for Post-Click Conversion Rate Estimation

Abstract: Accurate estimation of post-click conversion rate is critical for building recommender systems, which has long been confronted with sample selection bias and data sparsity issues. Methods in the Entire Space Multi-task Model (ESMM) family leverage the sequential pattern of user actions, \ie $impression\rightarrow click \rightarrow conversion$ to address data sparsity issue. However, they still fail to ensure the unbiasedness of CVR estimates. In this paper, we theoretically demonstrate that ESMM suffers from the following two problems: (1) Inherent Estimation Bias (IEB) for CVR estimation, where the CVR estimate is inherently higher than the ground truth; (2) Potential Independence Priority (PIP) for CTCVR estimation, where ESMM might overlook the causality from click to conversion. To this end, we devise a principled approach named Entire Space Counterfactual Multi-task Modelling (ESCM$^2$), which employs a counterfactual risk miminizer as a regularizer in ESMM to address both IEB and PIP issues simultaneously. Extensive experiments on offline datasets and online environments demonstrate that our proposed ESCM$^2$ can largely mitigate the inherent IEB and PIP issues and achieve better performance than baseline models.

Metadata-Induced Contrastive Learning for Zero-Shot Multi-Label Text Classification

Abstract: Large-scale multi-label text classification (LMTC) aims to associate a document with its relevant labels from a large candidate set. Most existing LMTC approaches rely on massive human-annotated training data, which are often costly to obtain and suffer from a long-tailed label distribution (i.e., many labels occur only a few times in the training set). In this paper, we study LMTC under the zero-shot setting, which does not require any annotated documents with labels and only relies on label surface names and descriptions. To train a classifier that calculates the similarity score between a document and a label, we propose a novel metadata-induced contrastive learning (MICoL) method. Different from previous text-based contrastive learning techniques, MICoL exploits document metadata (e.g., authors, venues, and references of research papers), which are widely available on the Web, to derive similar document–document pairs. Experimental results on two large-scale datasets show that: (1) MICoL significantly outperforms strong zero-shot text classification and contrastive learning baselines; (2) MICoL is on par with the state-of-the-art supervised metadata-aware LMTC method trained on 10K–200K labeled documents; and (3) MICoL tends to predict more infrequent labels than supervised methods, thus alleviates the deteriorated performance on long-tailed labels.

Beyond G1/S regulation: how cell size homeostasis is tightly controlled throughout the cell cycle?

Abstract: Proliferating cells require control mechanisms to counteract stochastic noise and stabilize the cell mass distribution in a population. It is widely believed that size-dependent timing of the G1/S transition is the predominant control process in mammalian cells. However, a model based only on such a checkpoint cannot explain why cell lines with deficient G1/S control are still able to maintain a stable cell mass distribution or how cell mass is maintained throughout the subsequent phases of the cell cycle. To answer such questions, we used a recently developed form of Quantitative Phase Microscopy (ceQPM), which provides much-improved accuracy of individual cell mass measurement, to investigate the factors contributing to cell mass homeostasis. We find that cell mass homeostasis is robustly maintained, despite disruptions of the normal G1/S transition or cell growth rate. The coefficient of variation in mass, nevertheless declines well after the G1/S transition and throughout the cell cycle in both transformed and non-transformed cells. Furthermore, the cell growth rate responds to cell mass in different ways in different cell types. Growth rate modulation is conveyed by mTORC1-dependent and mTORC1-independent processes, which are independently regulated. Slightly reduced mass accumulation, below exponential growth can effectively reduce cell mass variation in a population. Both size-dependent cell cycle regulation and size-dependent growth rate modulation contribute to cell mass homeostasis. Together they form a compensatory network that strictly controls the coefficient of variation of cell mass in a population. These findings opens new avenues for the discovery of underlying molecular mechanisms. Furthermore, understanding feedback directly on cell growth may provide insights into fundamental principles of cell size control in normal proliferating and terminally differentiated cells, as well as cells in pathological circumstances.

The Interaction Between N6-Methyladenosine Modification and Non-Coding RNAs in Gastrointestinal Tract Cancers

Abstract: N6-methyladenosine (m6A) is the most common epigenetic modification of eukaryotic RNA, which can participate in the growth and development of the body and a variety of physiological and disease processes by affecting the splicing, processing, localization, transport, translation, and degradation of RNA. Increasing evidence shows that non-coding RNAs, particularly microRNA, long non-coding RNA, and circular RNA, can also regulate the RNA m6A modification process by affecting the expression of m6A-related enzymes. The interaction between m6A modification and non-coding RNAs provides a new perspective for the exploration of the potential mechanism of tumor genesis and development. In this review, we summarize the potential mechanisms and effects of m6A and non-coding RNAs in gastrointestinal tract cancers.

TaxoCom: Topic Taxonomy Completion with Hierarchical Discovery of Novel Topic Clusters

Abstract: Topic taxonomies, which represent the latent topic (or category) structure of document collections, provide valuable knowledge of contents in many applications such as web search and information filtering. Recently, several unsupervised methods have been developed to automatically construct the topic taxonomy from a text corpus, but it is challenging to generate the desired taxonomy without any prior knowledge. In this paper, we study how to leverage the partial (or incomplete) information about the topic structure as guidance to find out the complete topic taxonomy. We propose a novel framework for topic taxonomy completion, named TaxoCom, which recursively expands the topic taxonomy by discovering novel sub-topic clusters of terms and documents. To effectively identify novel topics within a hierarchical topic structure, TaxoCom devises its embedding and clustering techniques to be closely-linked with each other: (i) locally discriminative embedding optimizes the text embedding space to be discriminative among known (i.e., given) sub-topics, and (ii) novelty adaptive clustering assigns terms into either one of the known sub-topics or novel sub-topics. Our comprehensive experiments on two real-world datasets demonstrate that TaxoCom not only generates the high-quality topic taxonomy in terms of term coherency and topic coverage but also outperforms all other baselines for a downstream task.

Teleostean fishes may have developed an efficient Na+ uptake for adaptation to the freshwater system

Abstract: Understanding Na+ uptake mechanisms in vertebrates has been a research priority since vertebrate ancestors were thought to originate from hyperosmotic marine habitats to the hypoosmotic freshwater system. Given the evolutionary success of osmoregulator teleosts, these freshwater conquerors from the marine habitats are reasonably considered to develop the traits of absorbing Na+ from the Na+-poor circumstances for ionic homeostasis. However, in teleosts, the loss of epithelial Na+ channel (ENaC) has long been a mystery and an issue under debate in the evolution of vertebrates. In this study, we evaluate the idea that energetic efficiency in teleosts may have been improved by selection for ENaC loss and an evolved energy-saving alternative, the Na+/H+ exchangers (NHE3)-mediated Na+ uptake/NH4 + excretion machinery. The present study approaches this question from the lamprey, a pioneer invader of freshwater habitats, initially developed ENaC-mediated Na+ uptake driven by energy-consuming apical H+-ATPase (VHA) in the gills, similar to amphibian skin and external gills. Later, teleosts may have intensified ammonotelism to generate larger NH4 + outward gradients that facilitate NHE3-mediated Na+ uptake against an unfavorable Na+ gradient in freshwater without consuming additional ATP. Therefore, this study provides a fresh starting point for expanding our understanding of vertebrate ion regulation and environmental adaptation within the framework of the energy constraint concept.

In Vivo Functional Assay in Fish Gills: Exploring Branchial Acid-Excreting Mechanisms in Zebrafish

Abstract: Molecular and physiological analyses in ionoregulatory organs (e.g., adult gills and embryonic skin) are essential for studying fish ion regulation. Recent progress in the molecular physiology of fish ion regulation was mostly obtained in embryonic skin; however, studies of ion regulation in adult gills are still elusive and limited because there are no direct methods for in vivo functional assays in the gills. The present study applied the scanning ion-selective electrode technique (SIET) in adult gills to investigate branchial H+-excreting functions in vivo. We removed the opercula from zebrafish and then performed long-term acid acclimation experiments. The results of Western blot and immunofluorescence showed that the protein expression of H+-ATPase (HA) and the number of H+-ATPase-rich ionocytes were increased under acidic situations. The SIET results proved that the H+ excretion capacity is indeed enhanced in the gills acclimated to acidic water. In addition, both HA and Na+/H+ exchanger (Nhe) inhibitors suppressed the branchial H+ excretion capacity, suggesting that H+ is excreted in association with HA and Nhe in zebrafish gills. These results demonstrate that SIET is effective for in vivo detection in fish gills, representing a breakthrough approach for studying the molecular physiology of fish ion regulation.

Illite K-Ar and (U-Th)/He low-temperature thermochronology reveal onset timing of Yadong-Gulu rift in southern Tibetan Plateau

Abstract: Determining the timing of E-W extension across the NS-trending rifts in southern Tibet is key to test the mechanical models of the latest evolution in the collision between India and Asia. We focus on the southern half of the largest of the seven main rifts, the Yadong-Gulu rift (YGR), which, despite being the focus of numerous studies thanks to its easy access, still lacks direct time constraints. Using illite K-Ar ages of fault gouge from the active Yadong normal fault of the YGR, we directly constrain its onset timing at 9 ± 1 Ma. (U-Th)/He dating of the footwall leucogranite reveals a rapid exhumation of the southern YGR since ∼9 Ma, attesting to its onset activity. Such timing is similar to that estimated for the northern half of the YGR at 8 ± 1 Ma, suggesting that the entire YGR formed at approximately the same time. Our synthesis of published initiation ages of the other main rifts in southern Tibet shows that they mostly fall between ∼23 and 8 Ma, suggesting a clear spatial and temporal pattern of old initiation ages to the west and young to the east. In this case, the formation of rifts in southern Tibet is unlikely caused by slab tearing of the underthruting Indian plate or orogenic collapse. Our study supports that E-W extension in Tibetan Plateau is triggered by a combination of eastward propagation of the Karakorum-Jiali fault zone and divergent thrusting along the curved Himalayan arc.

DSRRTracker: Dynamic Search Region Refinement for Attention-based Siamese Multi-Object Tracking

Abstract: Many multi-object tracking (MOT) methods follow the framework of"tracking by detection", which associates the target objects-of-interest based on the detection results. However, due to the separate models for detection and association, the tracking results are not optimal.Moreover, the speed is limited by some cumbersome association methods to achieve high tracking performance. In this work, we propose an end-to-end MOT method, with a Gaussian filter-inspired dynamic search region refinement module to dynamically filter and refine the search region by considering both the template information from the past frames and the detection results from the current frame with little computational burden, and a lightweight attention-based tracking head to achieve the effective fine-grained instance association. Extensive experiments and ablation study on MOT17 and MOT20 datasets demonstrate that our method can achieve the state-of-the-art performance with reasonable speed.

Syntectonic Sediment Recycling Controls Eolian Deposition in Eastern Asia Since ∼8 Ma

Abstract: Global cooling and/or Tibetan Plateau uplift have long been regarded as the principal drivers of late Cenozoic central Asian aridification and the resulting widespread accumulation of eolian deposits in eastern Asia. However, these two factors are unable to form large source areas of fine‐grained sediments enhancing eolian deposition synchronously from northern Tibet to North Pacific. Here we provide magnetostratigraphic and detrital apatite fission‐track evidence for a major sediment recycling event in northern Tibet at ∼8 Ma, coeval with a sudden increase in eolian deposition, which we ascribe to syntectonic erosion of uplifted friable fluvio‐lacustrine sediments and selective entrainment by the westerly winds during basin deformation. Our results emphasize the importance of widespread and persistent occurrence of fine‐grained sediments along the pathway of westerlies to produce voluminous dust deposits. These findings suggest that the onset of eolian deposition may not be directly related to global cooling or uplift of mountain ranges.

Target capture sequencing reveals a monoclonal outbreak of respiratory syncytial virus B infections among adult hematologic patients

Abstract: Abstract Background Respiratory syncytial virus (RSV) causes community-acquired respiratory tract infections during winter. However, outbreaks in hospitals also occur repeatedly. In particular, patients with hematologic malignancies are at an increased risk for a severe and potentially fatal course of RSV infection. Here we present the investigation of an RSV outbreak in a hematology ward for adults following the ORION statement. Methods An epidemiologic and molecular outbreak analysis was performed. We developed and employed a minimal oligonucleotide probe set in target capture probe sequencing that allows cost-effective RSV-A or -B capturing to reconstruct RSV genomes from clinical samples. Results Four adult patients were involved in the outbreak caused by RSV-B in March 2019. The enforcement of the pre-existing infection control measures by effective training of hospital staff contributed to a successful containment. PCR-based RSV screening on the ward enabled early detection of new cases and rapid isolation measures. The molecular analysis demonstrated that the outbreak sequences were highly related and distinct to other RSV-B strains circulating at the same time. Conclusions A multimodal infection control concept is essential for the timely detection and control of RSV outbreaks in patients with hematological disease. Among other measures, preventive screening for respiratory viruses is recommended. Furthermore, the integration of conventional and molecular epidemiology, such as whole-genome sequencing and variant calling, significantly contributes to the understanding of transmission pathways. Based on this, appropriate conclusions can be drawn for targeted prevention measures that have prepared us for the COVID-19 pandemic beyond the RSV approach described here.

New mechanism of FA in composted sludge inducing Cu fixation on Albite in open-pit mine soil.

Abstract: Fulvic acid (FA), typical organic matter derived from humification process in composted sludge, possesses the potential to remediate mine soils contaminated by heavy metals. To understand the cooper (Cu) immobilizing process in open-pit mine soil induced by FA, changes of Cu speciation in mixture of open-pit mine soil and composted sludge was tracked over 180 days. It was observed that the organic-bound and residual fraction of Cu increased dramatically with the corresponding decrease of Fe/Mn oxide-bound Cu in the first 60 days, then the organic-bound fraction decreased to about its initial proportion during 60-120 days, while residual fraction still increased, and the proportion of residual Cu accounted for over 85% and became stable after 120 days. To reveal the mechanism of FA inducing Cu fixation on Albite which is the main phase of soil primary ore, two groups of Cu adsorption experiments with and without FA were designed. With the addition of FA, the adsorption capacity of Cu by Albite increased by 1.55 times and the content of residual Cu in Albite increased by 7.7 times. It was found that the Cu absorbed in smaller Albite particle induced by FA formed a secondary mineral--Chrysocolla, causing increase of residual fraction of Cu. These results revealed the mechanism: FA was absorbed on the surface of Albite after complexing with Cu ions in the solution, and then it induced Cu into the interlayer and pore channels of Albite. The Cu in the Albite was immobilized by forming Chrysocolla finally.

Target capture sequencing reveals a monoclonal outbreak of respiratory syncytial virus B infections among adult hematologic patients

Abstract: Abstract Background Respiratory syncytial virus (RSV) causes community-acquired respiratory tract infections during winter. However, outbreaks in hospitals also occur repeatedly. In particular, patients with hematologic malignancies are at an increased risk for a severe and potentially fatal course of RSV infection. Here we present the investigation of an RSV outbreak in a hematology ward for adults following the ORION statement. Methods An epidemiologic and molecular outbreak analysis was performed. We developed and employed a minimal oligonucleotide probe set in target capture probe sequencing that allows cost-effective RSV-A or -B capturing to reconstruct RSV genomes from clinical samples. Results Four adult patients were involved in the outbreak caused by RSV-B in March 2019. The enforcement of the pre-existing infection control measures by effective training of hospital staff contributed to a successful containment. PCR-based RSV screening on the ward enabled early detection of new cases and rapid isolation measures. The molecular analysis demonstrated that the outbreak sequences were highly related and distinct to other RSV-B strains circulating at the same time. Conclusions A multimodal infection control concept is essential for the timely detection and control of RSV outbreaks in patients with hematological disease. Among other measures, preventive screening for respiratory viruses is recommended. Furthermore, the integration of conventional and molecular epidemiology, such as whole-genome sequencing and variant calling, significantly contributes to the understanding of transmission pathways. Based on this, appropriate conclusions can be drawn for targeted prevention measures that have prepared us for the COVID-19 pandemic beyond the RSV approach described here.

Bone metastasis risk and prognosis assessment models for kidney cancer based on machine learning

Abstract: Background Bone metastasis is a common adverse event in kidney cancer, often resulting in poor survival. However, tools for predicting KCBM and assessing survival after KCBM have not performed well. Methods The study uses machine learning to build models for assessing kidney cancer bone metastasis risk, prognosis, and performance evaluation. We selected 71,414 kidney cancer patients from SEER database between 2010 and 2016. Additionally, 963 patients with kidney cancer from an independent medical center were chosen to validate the performance. In the next step, eight different machine learning methods were applied to develop KCBM diagnosis and prognosis models while the risk factors were identified from univariate and multivariate logistic regression and the prognosis factors were analyzed through Kaplan-Meier survival curve and Cox proportional hazards regression. The performance of the models was compared with current models, including the logistic regression model and the AJCC TNM staging model, applying receiver operating characteristics, decision curve analysis, and the calculation of accuracy and sensitivity in both internal and independent external cohorts. Results Our prognosis model achieved an AUC of 0.8269 (95%CI: 0.8083–0.8425) in the internal validation cohort and 0.9123 (95%CI: 0.8979–0.9261) in the external validation cohort. In addition, we tested the performance of the extreme gradient boosting model through decision curve analysis curve, Precision-Recall curve, and Brier score and two models exhibited excellent performance. Conclusion Our developed models can accurately predict the risk and prognosis of KCBM and contribute to helping improve decision-making.

Escm2

Abstract: Accurate estimation of post-click conversion rate is critical for building recommender systems, which has long been confronted with sample selection bias and data sparsity issues. Methods in the Entire Space Multi-task Model (ESMM) family leverage the sequential pattern of user actions, i.e. $impression\rightarrow click \rightarrow conversion$ to address data sparsity issue. However, they still fail to ensure the unbiasedness of CVR estimates. In this paper, we theoretically demonstrate that ESMM suffers from the following two problems: (1) Inherent Estimation Bias (IEB), where the estimated CVR of ESMM is inherently higher than the ground truth; (2) Potential Independence Priority (PIP) for CTCVR estimation, where there is a risk that the ESMM overlooks the causality from click to conversion. To this end, we devise a principled approach named Entire Space Counterfactual Multi-task Modelling (ESCM$^2$), which employs a counterfactual risk miminizer as a regularizer in ESMM to address both IEB and PIP issues simultaneously. Extensive experiments on offline datasets and online environments demonstrate that our proposed ESCM$^2$ can largely mitigate the inherent IEB and PIP issues and achieve better performance than baseline models.

Analysis of N6-Methyladenosine Methylome in Adenocarcinoma of Esophagogastric Junction

Abstract: Background: From previous studies, we found that there are more than 100 types of RNA modifications in RNA molecules. m6A methylation is the most common. The incidence rate of adenocarcinoma of the esophagogastric junction (AEG) at home and abroad has increased faster than that of stomach cancer at other sites in recent years. Here, we systematically analyze the modification pattern of m6A mRNA in adenocarcinoma at the esophagogastric junction. Methods: m6A sequencing, RNA sequencing, and bioinformatics analysis were used to describe the m6A modification pattern in adenocarcinoma and normal tissues at the esophagogastric junction. Results: In AEG samples, a total of 4,775 new m6A peaks appeared, and 3,054 peaks disappeared. The unique m6A-related genes in AEG are related to cancer-related pathways. There are hypermethylated or hypomethylated m6A peaks in AEG in differentially expressed mRNA transcripts. Conclusion: This study preliminarily constructed the first m6A full transcriptome map of human AEG. This has a guiding role in revealing the mechanism of m6A-mediated gene expression regulation.

Respiratory Syncytial Virus Two-Step Infection Screen Reveals Inhibitors of Early and Late Life Cycle Stages

Abstract: Human respiratory syncytial virus (hRSV) infection is a leading cause of severe respiratory tract infections. Effective, directly acting antivirals against hRSV are not available. ABSTRACT Human respiratory syncytial virus (hRSV) infection is a leading cause of severe respiratory tract infections. Effective, directly acting antivirals against hRSV are not available. We aimed to discover new and chemically diverse candidates to enrich the hRSV drug development pipeline. We used a two-step screen that interrogates compound efficacy after primary infection and a consecutive virus passaging. We resynthesized selected hit molecules and profiled their activities with hRSV lentiviral pseudotype cell entry, replicon, and time-of-addition assays. The breadth of antiviral activity was tested against recent RSV clinical strains and human coronavirus (hCoV-229E), and in pseudotype-based entry assays with non-RSV viruses. Screening 6,048 molecules, we identified 23 primary candidates, of which 13 preferentially scored in the first and 10 in the second rounds of infection, respectively. Two of these molecules inhibited hRSV cell entry and selected for F protein resistance within the fusion peptide. One molecule inhibited transcription/replication in hRSV replicon assays, did not select for phenotypic hRSV resistance and was active against non-hRSV viruses, including hCoV-229E. One compound, identified in the second round of infection, did not measurably inhibit hRSV cell entry or replication/transcription. It selected for two coding mutations in the G protein and was highly active in differentiated BCi-NS1.1 lung cells. In conclusion, we identified four new hRSV inhibitor candidates with different modes of action. Our findings build an interesting platform for medicinal chemistry-guided derivatization approaches followed by deeper phenotypical characterization in vitro and in vivo with the aim of developing highly potent hRSV drugs.

Boosting Source-free Domain Adaptation via Confidence-based Subsets Feature Alignment

Abstract: Source-free Domain Adaptation (SFDA) aims to adapt a model trained on a given (source) environment to the new (target) environment, without directly accessing the source data. Due to the lack of labeled source data, it is often difficult for SFDA methods to provide reliable class representations for the target data. To overcome this issue, we propose the idea of Confidence-based Subsets Feature Alignment (CSFA). CSFA divides the target data into two subsets: confident subset that consists of samples having low entropy class predictions from the source model, and non-confident subset with samples that do not. By using the pseudo-labels from the confident subset, we can frame the original SFDA problem as a Universal Domain Adaptation (UniDA) problem, and provide reliable class representations for the target data by aligning feature distributions of the two subsets. Specifically, we propose a multi-task framework that simultaneously applies a standard SFDA algorithm in combination with a UniDA-inspired algorithm, which further infuses class representations into the adaption process. We evaluate the proposed method on a wide range of cross-domain object recognition tasks and achieve higher or comparable accuracy compared to existing SFDA methods. Ablation studies are conducted to verify the effectiveness of the proposed method.

Self-Supervised Cross-Video Temporal Learning for Unsupervised Video Domain Adaptation

Abstract: We address the task of unsupervised domain adaptation (UDA) for videos with self-supervised learning. While UDA for images is a widely studied problem, UDA for videos is relatively unexplored. In this paper, we propose a novel self-supervised loss for the task of video UDA. The method is motivated by inverted reasoning. Many works on video classification have shown success with representations based on events in videos, e.g., ‘reaching’, ‘picking’, and ‘drinking’ events for ‘drinking coffee’. We argue that if we have event-based representations, we should be able to predict the relative distances between clips in videos. Inverting that, we propose a self-supervised task to predict the difference of the distance between two clips from the source video and the distance between two clips from the target video. We hope that such a task would encourage learning event-based representations of the videos, which is known to be beneficial for classification. Since we predict the difference of clip distances between clips from source videos and target videos, we ‘tie’ the two domains and expect to achieve well-adapted representations. We combine this purely self-supervised loss and the source classification loss to learn the model parameters. We give extensive empirical results on challenging video UDA benchmarks, i.e., UCF-HMDB and EPIC-Kitchens. The presented qualitative and quantitative results support our motivations and method.

基于原位焦耳热处理的自粘附柔性表面肌电电极阵列设计与制备研究

Abstract: 随着柔性电子的发展，表皮电子在人机交互中的作用越来越突出。然而，柔性器件的稳定性和可靠性仍然是其得到广泛应用的一个挑战。本文设计一种以聚乙烯醇(PVA)为衬底，金(Au)为电极层的表面肌电(sEMG)电极阵列，通过原位焦耳加热方法对有源区域PVA衬底进行精确热处理调控，提高其结晶度和模量，提升有源区域电极的稳定性。基于该方法制备的sEMG电极阵列实现了与皮肤的共形贴附，可稳定采集8小时以上的高信噪比(达到21.3 dB)的肌电信号，进一步开展了基于sEMG信号的手势识别应用研究，识别率可达99.27%。以上结果表明，本研究提出的sEMG电极阵列制备方法简单高效，可长时间稳定工作，有望在基于sEMG信号的人机交互领域中得到应用。