A5022571158- Underwater Acoustics Research
- Underwater Vehicles and Communication Systems
- Glaucoma and retinal disorders
- Retinal Diseases and Treatments
- Blind Source Separation Techniques
- Corneal surgery and disorders
- Manufacturing Process and Optimization
- Video Surveillance and Tracking Methods
- Domain Adaptation and Few-Shot Learning
- Digital Transformation in Industry
- Geophysical Methods and Applications
- Collaboration in agile enterprises
- Speech and Audio Processing
- Anomaly Detection Techniques and Applications
- Intraocular Surgery and Lenses
- Hydraulic Fracturing and Reservoir Analysis
- Autonomous Vehicle Technology and Safety
- Advanced Image and Video Retrieval Techniques
- Business Process Modeling and Analysis
- Distributed and Parallel Computing Systems
- Advanced Vision and Imaging
- Advanced Neural Network Applications
- Flexible and Reconfigurable Manufacturing Systems
- Sperm and Testicular Function
- Human Pose and Action Recognition
Chinese Academy of Sciences
2019-2025
Institute of Acoustics
2019-2025
University of Chinese Academy of Sciences
2020-2025
Shaoxing Second Hospital
2025
China Telecom
2023-2024
China Telecom (China)
2023-2024
Beihang University
2020-2024
Second Affiliated Hospital of Xi'an Jiaotong University
2024
Jiangnan University
2024
Lake Forest Academy
2024
Deep classifiers have achieved great success in visual recognition. However, real-world data is long-tailed by nature, leading to the mismatch between training and testing distributions. In this paper, we show that Softmax function, though used most classification tasks, gives a biased gradient estimation under setup. This paper presents Balanced Softmax, an elegant unbiased extension of accommodate label distribution shift testing. Theoretically, derive generalization bound for multiclass...
Data imbalance exists ubiquitously in real-world visual regressions, e.g., age estimation and pose estimation, hurting the model's generalizability fairness. Thus, im-balanced regression gains increasing research attention recently. Compared to imbalanced classification, focuses on continuous labels, which can be boundless high-dimensional hence more challenging. In this work, we identify that widely used Mean Square Error (MSE) loss function ineffective regression. We revisit MSE from a...
SummaryBackgroundThe outbreak of COVID-19 has led to international concern. We aimed establish an effective screening strategy in Shanghai, China, aid early identification patients with COVID-19.MethodsWe did a multicentre, observational cohort study fever clinics 25 hospitals 16 districts Shanghai. All visiting the within period were included. A for was presented and then suspected cases monitored analysed until they confirmed as or excluded. Logistic regression used determine risk factors...
With the soil constitutive model developed for dynamic large deformation of soft soil, a finite element was constructed to simulate soil–underground structure static and coupling interaction system, in which contact between underground damage reinforced concrete were considered. The seismic performance mechanism subway station different foundations investigated evaluated detail. It proven that would change with type foundation peak acceleration input ground motion, can be by coefficient...
Antibiotics are an important pharmaceutical class excessively used by humans. Its presence in the soil can impact plant growth and induce antibiotic resistance. This research studies effect of sulfamethoxazole (SMX) on growth, rhizosphere bacteria composition, resistance genes. Two sets vegetables (basil, cilantro, spinach) were treated separately with water SMX solution. The data samples collected analyzed. results revealed that increased spinach leaf length (34.0%) while having no...
Epithelial-to-mesenchymal transition (EMT) has profound impacts on cancer progression and also drug resistance, including epidermal growth factor receptor tyrosine kinase inhibitors (EGFR-TKIs). Nowadays, there is still no predictive biomarker identified for the use of EGFR-TKIs in non-small cell lung (NSCLC) patients with wild-type EGFR. To clarify role EMT phenotype as a marker EGFR-TKI, we performed retrospective study 202 stage IV or recurrent NSCLC receiving gefitinib erlotinib therapy...
Underwater acoustic target recognition is an intractable task due to the complex source characteristics and sound propagation patterns. Limited by insufficient data narrow information perspective, models based on deep learning seem far from satisfactory in practical underwater scenarios. Although signals are severely influenced distance, channel depth, or other factors, annotations of relevant often nonuniform, incomplete, hard use. In this work, proposal implement templates made up rich...
To systematically evaluate the effectiveness of respiratory training interventions in rehabilitation patients with rib fractures through a meta-analysis, aiming to provide robust evidence for clinical practice. A comprehensive search was conducted multiple databases (Pubmed, Embase, Web Science, Cochrane Central, CNKI, Wanfang Data, and CSTJ) up August 15, 2024, identify relevant randomized controlled trials (RCTs). Eligible studies were those that compared RTIs plus conventional treatment...
Underwater passive acoustic recognition, which focuses on classifying targets based ship-radiated noise, is a key challenge in underwater acoustics. Deep learning-based methods have gained popularity recent years because of their strong performance. However, these often fail to generalize well real-world scenarios. This work reveals one underlying challenge: the characteristics noise are influenced by factors such as vessel structures and propulsion systems. Although vessels same type may...
Few studies have investigated the epidemiology of secondary glaucoma (SG) in China. This study analyzed current profile admitted patients with SG at largest ophthalmic center China.SG cases were retrieved retrospectively by International Classification Disease-10 clinical coding data from January 2010 to December 2019. Demographic data, etiologies, and management analyzed. The ratios etiologies compared previously reported other regions. between 2014 group 2015 2019 compared.A total 11,730...
Image- and video-based 3D human recovery (i.e., pose shape estimation) have achieved substantial progress. However, due to the prohibitive cost of motion capture, existing datasets are often limited in scale diversity. In this work, we obtain massive sequences by playing video game with automatically annotated ground truths. Specifically, contribute GTA-Human, a large-scale dataset generated GTA-V engine, featuring highly diverse set subjects, actions, scenarios. More importantly, study use...