A5100344706- Time Series Analysis and Forecasting
- Multimodal Machine Learning Applications
- Advanced Clustering Algorithms Research
- Anomaly Detection Techniques and Applications
- Domain Adaptation and Few-Shot Learning
- COVID-19 diagnosis using AI
- Face and Expression Recognition
- Data Stream Mining Techniques
- Advanced Image and Video Retrieval Techniques
- Robotics and Sensor-Based Localization
- Machine Learning in Healthcare
- Horticultural and Viticultural Research
- Berry genetics and cultivation research
- Advanced Neural Network Applications
- Plant Reproductive Biology
- Context-Aware Activity Recognition Systems
- Embedded Systems and FPGA Design
- Research on scale insects
- Frailty in Older Adults
- COVID-19 Pandemic Impacts
- Heart Rate Variability and Autonomic Control
- Chronic Obstructive Pulmonary Disease (COPD) Research
- Respiratory Support and Mechanisms
- Advanced Sensor and Control Systems
- Bayesian Methods and Mixture Models
Tencent (China)
2023
University of Missouri
2018-2023
Honghu Hospital of Traditional Chinese Medicine
2022
Hanshan Normal University
2012
Anshan Normal University
2012
The detector-free feature matching approaches are currently attracting great attention thanks to their excellent performance. However, these methods still struggle at large-scale and viewpoint variations, due the geometric inconsistency resulting from application of mutual nearest neighbour criterion (i.e., one-to-one assignment) in patch-level matching. Accordingly, we in-troduce AdaMatcher, which first accomplishes correlation co-visible area estimation through an elaborate interaction...
Recent Few-Shot Learning (FSL) methods put emphasis on generating a discriminative embedding features to precisely measure the similarity between support and query sets. Current CNN-based cross-attention approaches generate representations via enhancing mutually semantic similar regions of pairs. However, it suffers from two problems: CNN structure produces inaccurate attention map based local features, backgrounds cause distraction. To alleviate these problems, we design novel SpatialFormer...
Abstract Purpose Low serum creatinine/cystatin C ratio (CCR) is associated with unfavorable characteristics in patients chronic obstructive pulmonary disease (COPD); however, the relationship between CCR and in-hospital mortality of acute exacerbation COPD (AECOPD) unexplored. Our objective was to assess value for predicting hospitalized AECOPD. Methods Patients AECOPD ( n = 597) were retrospectively enrolled. Patient’s clinical laboratory tests, including cystatin creatinine, reviewed. The...
The Possibilistic C-Means (PCM) was developed as an extension of the Fuzzy (FCM) by abandoning membership sum-to-one constraint. In PCM, each cluster is independent other clusters, and can be processed separately. Thus, Sequential One-Means (SP1M) proposed to find clusters sequentially running P1M C times. One critical problem in both PCM SP1M how determine parameter η. Means with Adaptive Eta (SP1M-AE) allow η change during iterations. this paper, we introduce a new dynamic adaption...
The rapid aging of the population worldwide requires increased attention from healthcare providers and entire society. For elderly to live independently, many health issues related old age, such as frailty risk falling, need monitoring. When monitoring daily routines for older adults, it is desirable detect early signs changes before serious events, hospitalizations, happen so that timely adequate preventive care may be provided. By deploying multi-sensor systems in homes elderly, we can...
Examining most streaming clustering algorithms leads to the understanding that they are actually incremental classification models. They model existing and newly discovered structures via summary information we call footprints. Incoming data is normally assigned a crisp label (into one of structures) structure's footprint incrementally updated. There no reason these assignments need be crisp. In this paper, propose new algorithm uses Neural Gas prototypes as footprints produces possibilistic...
Data stream processing has gained much attention lately, in the era of big data. Streaming clustering is an effective tool to recognize normal baseline and detect outliers sequentially presented Perhaps more importantly would be ability predict that incoming data indicates movement towards a likely anomaly. In this paper, Gaussian Mixture Model (GMM) employed represent different patterns stream. The Sequential Possibilistic One-Means (SP1M) used for initialization, incorporated into GMM...
Artificial neural networks are a dominant force in our modern era of data-driven artificial intelligence. The adaptive neuro fuzzy inference system (ANFIS) is network based on logic versus more traditional premise like convolution. Advantages ANFIS include the ability to encode and potentially understand machine learned information pursuit explainable, interpretable, ultimately trustworthy However, real-world data almost always imperfect, e.g., incomplete or noisy, not naturally robust....
Clustering has long been applied to the problem of image segmentation. Because spatial connectivity constraints, several approaches have proposed incorporate local consistency into segmentation by clustering. One popular method, fuzzy information c-means (FLICM) shown produce good results. Like cmeans (FCM) from which it is derived, FLICM requires that pixels "share" memberships across clusters, is, a pixel all clusters need sum one. The possibilistic (PCM) clustering was introduced relax...
The ability to explain the predictions of machine learning models has become increasingly important, especially in healthcare applications. Streaming clustering is an effective tool recognize normal baseline patterns and detect early signs changes data streams. However, many streaming algorithms are not designed users how made. In this paper, we extend a algorithm, sequential possibilistic Gaussian mixture model (SPGMM) for detection health change provide algorithm explainability results....
Blood pressure (BP) is a crucial indicator of cardiovascular health and provides essential information about the heart. Traditional cuff-based BP measurement with equipment like sphygmomanometers uncomfortable discontinued. Cuffless monitoring using photoplethysmography (PPG) signals has become an area interest in research community, particularly advent wearables such as Apple Watch Xiaomi Band. While most wearable PPG are obtained from wrist, widely used benchmark dataset, MIMIC collects...
Examining most streaming clustering algorithms leads to the understanding that they are actually incremental classification models. They model existing and newly discovered structures via summary information we call footprints. Incoming data is normally assigned a crisp label (into one of structures) structure's footprint incrementally updated. There no reason these assignments need be crisp. In this paper, propose new algorithm uses Neural Gas prototypes as footprints produces possibilistic...
The detector-free feature matching approaches are currently attracting great attention thanks to their excellent performance. However, these methods still struggle at large-scale and viewpoint variations, due the geometric inconsistency resulting from application of mutual nearest neighbour criterion (\ie, one-to-one assignment) in patch-level matching.Accordingly, we introduce AdaMatcher, which first accomplishes correlation co-visible area estimation through an elaborate interaction...
Recent Few-Shot Learning (FSL) methods put emphasis on generating a discriminative embedding features to precisely measure the similarity between support and query sets. Current CNN-based cross-attention approaches generate representations via enhancing mutually semantic similar regions of pairs. However, it suffers from two problems: CNN structure produces inaccurate attention map based local features, backgrounds cause distraction. To alleviate these problems, we design novel SpatialFormer...
Weak feature representation problem has influenced the performance of few-shot classification task for a long time. To alleviate this problem, recent researchers build connections between support and query instances through embedding patch features to generate discriminative representations. However, we observe that there exists semantic mismatches (foreground/ background) among these local patches, because location size target object are not fixed. What is worse, result in unreliable...
This paper presents the real-world smart-meter dataset and offers an analysis of solutions derived from Energy Prediction Technical Challenges, focusing primarily on two key competitions: IEEE Computational Intelligence Society (IEEE-CIS) Challenge Smart Meter data in 2020 (named EP) its follow-up challenge at International Conference Fuzzy Systems (FUZZ-IEEE) 2021 as XEP). These competitions focus accurate energy consumption forecasting importance interpretability understanding underlying...
The rapid increase of the older population has gained significant healthcare attention in recent years. Early detection illness might result better outcomes and reduced cost. use health monitoring sensors such, motion or bed depth ones, may be able to detect early sign but they are hard interpret. To enhance caretakers understanding sensor data we propose link it signs symptoms extracted from related electronic record (EHR). proposed methodology is based on detecting changes patterns using...
Few-Shot Learning (FSL) alleviates the data shortage challenge via embedding discriminative target-aware features among plenty seen (base) and few unseen (novel) labeled samples. Most feature modules in recent FSL methods are specially designed for corresponding learning tasks (e.g., classification, segmentation, object detection), which limits utility of features. To this end, we propose a light universal module named transformer-based Semantic Filter (tSF), can be applied different tasks....
Abstract Older adults have experienced greater isolation and mental health concerns during the COVID-19 pandemic. In long-term care (LTC) settings, residents been particularly impacted due to strict lockdown policies. Little is known about how these policies older adults. This study leveraged existing research with embedded sensors installed in LTC analyzed sensor data of (N=30) two months pre/post onset U.S. pandemic (1/13/20 3/13/20, 03/14/20 5/13/20). Data from three (bed sensors, depth...