A5108050442- Anomaly Detection Techniques and Applications
- AI in cancer detection
- Radiomics and Machine Learning in Medical Imaging
- Adversarial Robustness in Machine Learning
- Machine Learning and Data Classification
- Advanced Neural Network Applications
- Topic Modeling
- EEG and Brain-Computer Interfaces
- Digital Media Forensic Detection
- Speech Recognition and Synthesis
- Advanced Graph Neural Networks
- 3D Printing in Biomedical Research
- Time Series Analysis and Forecasting
- Human Pose and Action Recognition
- Face and Expression Recognition
- Cancer Cells and Metastasis
- Text and Document Classification Technologies
- Privacy-Preserving Technologies in Data
- Domain Adaptation and Few-Shot Learning
- Explainable Artificial Intelligence (XAI)
- COVID-19 diagnosis using AI
- Phonocardiography and Auscultation Techniques
- ECG Monitoring and Analysis
- Context-Aware Activity Recognition Systems
- Speech and Audio Processing
Tsinghua–Berkeley Shenzhen Institute
2020-2024
Tsinghua University
2015-2024
Huawei Technologies (China)
2023-2024
Lanzhou Jiaotong University
2024
Shaanxi Normal University
2024
Beijing Normal University
2021-2024
ShanghaiTech University
2024
Shanghai Medical College of Fudan University
2023
Fudan University Shanghai Cancer Center
2017-2023
Xi'an University of Architecture and Technology
2023
Due to the low contrast, blurry boundaries, and large amount of shadows in breast ultrasound (BUS) images, automatic tumor segmentation remains a challenging task. Deep learning provides solution this problem, since it can effectively extract representative features from lesions background BUS images.A novel method is proposed by combining dilated fully convolutional network (DFCN) with phase-based active contour (PBAC) model. The DFCN an improved neural convolution deeper layers, fewer...
Transplantation of mesenchymal stem cells (MSCs) holds promise to repair severe traumatic injuries. However, current transplantation practices limit the potential this technique, either by losing viable MSCs or reducing performance resident MSCs. Herein, we design a "bead-jet" printer, specialized for high-throughput intra-operative formulation and printing MSCs-laden Matrigel beads. We show that high-density encapsulation in beads is able augment MSC function, increasing proliferation,...
Synthetic mRNAs are rising rapidly as alternative therapeutic agents for delivery of proteins. However, the practical use synthetic has been restricted by their low cellular stability well poor protein production efficiency. The key roles poly(A) tail on mRNA biology inspire us to explore optimization sequence overcome aforementioned limitations. Here, systematic substitution non-A nucleotides in tails revealed that cytidine-containing can substantially enhance rate and duration both vitro...
Multi-view clustering (MVC) is a popular technique for improving performance using various data sources. However, existing methods primarily focus on acquiring consistent information while often neglecting the issue of redundancy across multiple views. This study presents new approach called Sufficient Multi-View Clustering (SUMVC) that examines multi-view framework from an information-theoretic standpoint. Our proposed method consists two parts. Firstly, we develop simple and reliable SCMVC...
Abstract Ferroelectric 2D van der Waals (vdW) layered materials are attracting increasing attention due to their potential applications in next‐generation nanoelectronics and in‐memory computing with polarization‐dependent functionalities. Despite the critical role of polarization governing ferroelectricity behaviors, its origin relation local structures vdW have not been fully elucidated so far. Here, intralayer sliding approximately six degrees within each quadruple‐layer prototype...
Abstract The circadian clock coordinates the daily rhythmicity of biological processes, and its dysregulation is associated with various human diseases. Despite direct targeting rhythmic genes by many prevalent World Health Organization (WHO) essential drugs, traditional approaches can't satisfy need explore multi‐timepoint drug administration strategies across a wide range drugs. Here, droplet‐engineered primary liver organoids (DPLOs) are generated characteristics in 4 days, developed...
Broiler chickens are traditionally weighed by steelyard or platform scale, which is time-consuming and labor-intensive. usually exhibit stress-related behavior during weighing. The 3D camera-based weighing system for broiler can only weigh the chicken in monitoring area. Usually, it makes poor weight prediction due to segmentation especially when flapping its wings. To solve these issues, we developed one simple low-cost with high stability accuracy. A validity value extraction method from...
Electrocardiograph (ECG) is one of the most critical physiological signals for arrhythmia diagnosis in clinical practice. In recent years, various algorithms based on deep learning have been proposed to solve heartbeat classification problem and achieved saturated accuracy intrapatient paradigm, but encountered performance degradation inter-patient paradigm due drastic variation ECG among different individuals. this paper, we propose a novel unsupervised domain adaptation scheme address...
Two-dimensional (2D) ferroelectric materials are promising substitutes of three-dimensional perovskite based ceramic materials. Yet most studies have been focused on the construction non-centrosymmetric 2D van der Waals and only a few constructed experimentally. Herein, we experimentally demonstrate co-existence voltage-tunable out-of-plane (OOP) in-plane (IP) ferroelectricity in few-layer InSe prepared by solution-processable method fabricate semiconductor channel transistors. The...
We introduce a reliable scheme for continuous-variable quantum key distribution (CV-QKD) by using orthogonal frequency division multiplexing (OFDM). As spectrally efficient technique, OFDM allows large number of closely spaced subcarrier signals used to carry data on several parallel streams or channels. place emphasis modulator impairments which would inevitably arise in the system and analyze how these affect OFDM-based CV-QKD system. Moreover, we also evaluate security asymptotic limit...
Deep Neural Networks (DNNs) usually work in an end-to-end manner. This makes the trained DNNs easy to use, but they remain ambiguous decision process for every test case. Unfortunately, interpretability of decisions is crucial some scenarios, such as medical or financial data mining and decision-making. In this paper, we propose a Tree-Network-Tree (TNT) learning framework explainable decision-making, where knowledge alternately transferred between tree model DNNs. Specifically, proposed TNT...
Runoff from the high-cold mountains area (HCMA) is most important water resource in arid zone, and its accurate forecasting key to scientific management of resources downstream basin. Constrained by scarcity meteorological hydrological stations HCMA inconsistency observed time series, simulation reconstruction mountain runoff have always been a focus cold region research. Based on observations Yurungkash Kalakash Rivers, upstream tributaries Hotan River northern slope Kunlun Mountains at...
Purpose Digital Breast Imaging Reporting and Data System (BI-RADS) features extracted from ultrasound images are essential in computer-aided diagnosis, prediction, prognosis of breast cancer. This study focuses on the reproducibility quantitative high-throughput BI-RADS presence variations due to different segmentation results, various machine models, multiple settings. Methods Dataset 1 consists 399 patients with invasive cancer is used as training set measure features, while dataset 2 138...
Due to the significant variability in waveforms and characteristics of ECG signals, developing fully automatic (i.e., requires no expert assistance) heartbeat classification algorithms with satisfactory performance on domain-shifted data remains challenging. In this letter, we propose a novel Mixup Asymmetric Tri-training (MIAT) method improve generalization ability classifiers domain shift scenarios. First, develop an ECG-based tri-branch CNN model, including one shared feature encoder...
Organoids are expected to function as effective human organ models for precision cancer studies and drug development. Currently, primary tissue-derived organoids, termed non-engineered organoids (NEOs), produced by manual pipetting or liquid handling that compromises organoid-organoid homogeneity organoid-tissue consistency. Droplet-based microfluidics enables automated organoid production with high homogeneity, consistency, a significantly improved spectrum. It takes advantage of...