A5101733681- Multimodal Machine Learning Applications
- Anomaly Detection Techniques and Applications
- Anesthesia and Neurotoxicity Research
- Domain Adaptation and Few-Shot Learning
- PI3K/AKT/mTOR signaling in cancer
- Human Pose and Action Recognition
- Video Surveillance and Tracking Methods
- Network Security and Intrusion Detection
- Advanced Neural Network Applications
- Intensive Care Unit Cognitive Disorders
- Advanced Computational Techniques and Applications
- Food Allergy and Anaphylaxis Research
- Visual Attention and Saliency Detection
- Cancer-related molecular mechanisms research
- Video Analysis and Summarization
- Advanced Image and Video Retrieval Techniques
- Generative Adversarial Networks and Image Synthesis
- Catalytic Processes in Materials Science
- TGF-β signaling in diseases
- Smart Grid and Power Systems
- Distributed and Parallel Computing Systems
- Advanced Vision and Imaging
- Asthma and respiratory diseases
- Advancements in Battery Materials
- Smart Grid Security and Resilience
Southern University of Science and Technology
2022-2025
Fudan University Shanghai Cancer Center
2022-2025
Tobacco Research Institute
2025
China Tobacco
2025
Zhongnan Hospital of Wuhan University
2015-2024
Wuhan University
2015-2024
Second Affiliated Hospital of Fujian Medical University
2019-2023
Fujian Medical University
2019-2023
China Three Gorges University
2009-2023
Third Xiangya Hospital
2023
Alveolar epithelial barrier dysfunction in response to inflammatory reaction contributes pulmonary edema acute lung injury(ALI).Irisin,a newly-found myokine,exerts the anti-inflammatory effects. This study aims investigate protective effects of irisin on lipopolysaccharide (LPS)-induced ALIin vivo and vitro, explore its underlying mechanism. Male SD rats A549 cells were divided into 4 groups: control group, LPS Irisin pretreated Irisin/Compound C(a special inhibitor AMPK)-treated group. The...
Recently, the Segment Anything Model (SAM) gains lots of attention rapidly due to its impressive segmentation performance on images. Regarding strong ability image and high interactivity with different prompts, we found that it performs poorly consistent in videos. Therefore, this report, propose Track (TAM), which achieves high-performance interactive tracking To be detailed, given a video sequence, only very little human participation, i.e., several clicks, people can track anything they...
Although mainstream unsupervised anomaly detection (AD) algorithms perform well in academic datasets, their performance is limited practical application due to the ideal experimental setting of clean training data. Training with noisy data an inevitable problem real-world but seldom discussed. This paper considers label-level noise image sensory for first time. To solve this problem, we proposed a memory-based AD method, SoftPatch, which efficiently denoises at patch level. Noise...
Objectives Although China has done a lot in strengthening the primary healthcare system, high turnover intention is still social problem to be reckoned with. The objective of this study explore mediating effect satisfaction between job burnout and intention. Design Cross-sectional study. Methods A cross-sectional was conducted make sense burnout, among workers central China. Structural equation modelling (SEM) performed with maximum likelihood estimation. mediation test carried out by using...
Hypoxic environment is a bottleneck of photodynamic therapy (PDT) in tumor treatment, as oxygen the critical substrate for photosensitivity reaction. Herein, sustained supply system based on cerium nanoparticles and hydrogel (GHCAC) was explored enhanced synergistic PDT gas therapy. Ceria were prepared drug carrier by self-assembly mediated hyaluronic acid (HA), targeting CD44 cervical cancer cells, followed photosensitizer l-arginine (l-Arg) loading. Then, GHCAC developed incorporating...
Due to their precise three-dimensional structures and tunable cavities, metallacages covalent cages have been widely used in various fields, such as catalysis, separation, sensing, biomedicine. As a result of the unique pattern directed synthesis modular assembly, can be modified many different ways construct stimuli-responsive for marker detection imaging biological applications. Additionally, agents drugs loaded into cavity achieve disease diagnosis treatment. This Review summarizes recent...
Neonatal heart undergoes metabolic conversion and cell cycle arrest preparing for the increased workload during adulthood. Herein, we report that neonatal ketone body elevation is a critical regulatory factor postnatal development. Through multiomics screening, found expression of 3-hydroxy-3-methylglutaryl-CoA synthase 2 (HMGCS2), rate-limiting enzyme ketogenesis, was transiently induced by colostrum in heart. Hmgcs2 knockout caused mitochondrial maturation defects. Meanwhile, development...
In the area of fewshot anomaly detection (FSAD), efficient visual feature plays an essential role in memory bank M-based methods. However, these methods do not account for relationship between and its rotated feature, drastically limiting performance. To push limits, we reveal that rotation-invariant property has a significant impact industrial-based FSAD. Specifically, utilize graph representation FSAD provide novel isometric invariant (VIIF) as measurement feature. As result, VIIF can...
Image-to-text generation aims to describe images using natural language. Recently, zero-shot image captioning based on pre-trained vision-language models (VLMs) and large language (LLMs) has made significant progress. However, we have observed empirically demonstrated that these methods are susceptible modality bias induced by LLMs tend generate descriptions containing objects (entities) do not actually exist in the but frequently appear during training (i.e., object hallucination). In this...
Abstract This study aimed to investigate the prognostic impact of lymphovascular and perineural invasions in patients with squamous cell carcinoma tongue who received surgery-based treatment at our institution between January 2013 December 2020. Patients were divided into four groups based on presence (P−/P +) (V−/V +): P–V−, P–V + , P V−, V . Log-rank Cox proportional hazard models used evaluate association /lymphovascular invasion overall survival (OS). Altogether, 127 included, 95...
Controllable image captioning is an emerging multimodal topic that aims to describe the with natural language following human purpose, $\textit{e.g.}$, looking at specified regions or telling in a particular text style. State-of-the-art methods are trained on annotated pairs of input controls and output captions. However, scarcity such well-annotated data largely limits their usability scalability for interactive AI systems. Leveraging unimodal instruction-following foundation models...
BackgroundSleep disorders can profoundly affect neurological function. We investigated changes in social and anxiety-related brain functional connectivity induced by sleep deprivation, the potential therapeutic effects of general anaesthetics propofol sevoflurane rats.MethodsTwelve-week-old male Sprague–Dawley rats were subjected to deprivation for 20 h per day (from 14:00 10:00 next day) 4 consecutive weeks. They free from remaining during which they received (40 mg kg–1 i.p.) or (2% 2 h)...
Unsupervised Anomaly Detection (UAD) with incremental training is crucial in industrial manufacturing, as unpredictable defects make obtaining sufficient labeled data infeasible. However, continual learning methods primarily rely on supervised annotations, while the application UAD limited due to absence of supervision. Current train separate models for different classes sequentially, leading catastrophic forgetting and a heavy computational burden. To address this issue, we introduce novel...
Toll-like receptor 4 (TLR4) is a crucial in neuroinflammation and apoptotic neuronal death, increasing evidences indicated that β2-microglobulin (B2M) thought to be major contributor age-related cognitive decline. In present study, we designed investigate the effects of TLR4 on B2M-induced Wild-type (WT) C57BL/6, knockout (TLR4 -KO) mice hippocampal neurons from two type were respectively divided into groups: (1) Veh group; (2) B2M-treated group. The behavioral responses measured using...
Existing methods for video-based person re- identification (ReID) mainly learn the appearance feature of a given pedestrian via extractor and aggregator. However, models would fail to large inter-class variance when different pedestrians have similar appearances. Considering that walking postures body proportions, we propose discriminative pose beyond video retrieval. Specifically, implement two-branch architecture separately feature, then concatenate them together inference. To first detect...
Severe hypotension resulting from anaphylactic shock may be refractory to epinephrine and impair cerebral oxygenation metabolism contributing morbidity mortality. Refractoriness could corrected by nitric oxide pathway inhibitors such as methylene blue.To compare the systemic regional (brain skeletal muscle) effects of blue given alone or in combination a rat model shock.Prospective laboratory study.University laboratory.Male Brown-Norway rats (n = 60).After sensitization induction ovalbumin,...
The effects of acute reduction in arterial blood pressure severe anaphylactic shock (AS) on cerebral flow are paramount importance to be investigated. We studied circulation and oxygenation a model AS compared it with pharmacologically induced hypotension similar magnitude. Anaphylactic was by 1 mg intravenous ovalbumin (OVA) sensitized rats. Rats were randomized three groups: (i) no resuscitation (OVA; n = 10) (ii) volume expansion (10 mL 10 min after OVA injection) (OVA + VE; 10); (iii)...
Semi-supervised learning is a challenging problem which aims to construct model by from limited labeled examples. Numerous methods for this task focus on utilizing the predictions of unlabeled instances consistency alone regularize networks. However, treating and data separately often leads discarding mass prior knowledge learned In paper, we propose novel method semi-supervised semantic segmentation named GuidedMix-Net, leveraging information guide instances. Specifically, GuidedMix-Net...
The recent rapid development of deep learning has laid a milestone in industrial Image Anomaly Detection (IAD). In this paper, we provide comprehensive review learning-based image anomaly detection techniques, from the perspectives neural network architectures, levels supervision, loss functions, metrics and datasets. addition, extract new setting manufacturing current IAD approaches under our proposed setting. Moreover, highlight several opening challenges for detection. merits downsides...