A5101852743- Fault Detection and Control Systems
- Radiomics and Machine Learning in Medical Imaging
- Ectopic Pregnancy Diagnosis and Management
- Advanced biosensing and bioanalysis techniques
- Digital Imaging for Blood Diseases
- Computational Drug Discovery Methods
- Advanced Memory and Neural Computing
- AI in cancer detection
- Radiation Dose and Imaging
- Image Retrieval and Classification Techniques
- Gestational Trophoblastic Disease Studies
- Appendicitis Diagnosis and Management
- Maternal and fetal healthcare
- Advanced Radiotherapy Techniques
- Machine Learning in Materials Science
- DNA and Biological Computing
- Advanced Image Processing Techniques
- Ocular Oncology and Treatments
- Thyroid Cancer Diagnosis and Treatment
- Nonmelanoma Skin Cancer Studies
- Imbalanced Data Classification Techniques
- Gene expression and cancer classification
- Advanced Vision and Imaging
- Data Mining Algorithms and Applications
- MRI in cancer diagnosis
East China Normal University
2022-2024
First Affiliated Hospital of Kunming Medical University
2019-2024
Kunming Medical University
2019-2024
Hunan University
2024
First Affiliated Hospital of Hunan University of Traditional Chinese Medicine
2021-2024
University of Würzburg
2023
University of California, San Diego
2015
Georgia State University
2012-2015
Cooper University Hospital
2006-2009
St. Jude Children's Research Hospital
2000-2004
This paper introduces a novel benchmark for efficient up-scaling as part of the NTIRE 2023 Real-Time Image Super-Resolution (RTSR) Challenge, which aimed to upscale images from 720p and 1080p resolution native 4K (×2 ×3 factors) in real-time on commercial GPUs. For this, we use new test set containing diverse ranging digital art gaming photography. We assessed methods devised SR by measuring their runtime, parameters, FLOPs, while ensuring minimum PSNR fidelity over Bicubic interpolation....
Abstract The present study aimed to construct prospective models for tumor grading of rectal carcinoma by using magnetic resonance (MR)‐based radiomics features. A set 118 patients with was analyzed. After imbalance‐adjustments the data Synthetic Minority Oversampling Technique (SMOTE), final randomized into training and validation at ratio 3:1. features were captured from manually segmented lesion imaging (MRI). most related selected random forest model calculating Gini importance initial...
Medical image segmentation demands the aggregation of global and local feature representations, posing a challenge for current methodologies in handling both long-range short-range interactions. Recently, vision mamba (ViM) models have emerged as promising solutions addressing model complexities by excelling iterations with linear complexity. However, existing ViM approaches overlook importance preserving dependencies directly flattening spatial tokens are constrained fixed scanning patterns...
Medical image segmentation demands the aggregation of global and local feature representations, posing a challenge for current methodologies in handling both long-range short-range interactions. Recently, vision mamba (ViM) models have emerged as promising solutions addressing model complexities by excelling iterations with linear complexity. However, existing ViM approaches overlook importance preserving dependencies directly flattening spatial tokens are constrained fixed scanning patterns...
Modeling of a new domain can be challenging due to scarce data and high-dimensionality. Transfer learning aims integrate the with knowledge about some related old domains, in order model better. This paper studies transfer for degenerate biological systems. Degeneracy refers phenomenon that structurally different elements system perform same/similar function or yield output. exits various systems contributes heterogeneity, complexity, robustness is models enabling such have been little...
DNA reactions are crucial in biology, synthetic and computing. Accurate prediction of thermodynamic kinetic parameters is vital for understanding molecular interactions designing functional DNA-based systems. Existing models have limitations due to simplifications approximations that may deviate from experimental measurements. In this study, we propose a quantum chemistry-based deep learning model enhance accuracy efficiency predicting reaction parameters. The integrates chemistry...
<title>Abstract</title> The computational challenges inherent to using digital pathology for prognosis derive from the fact that a typical gigapixel slide may consist of thousands image tiles and previous models lack capability on modeling crucial slide-level contextual information, thereby resulting in suboptimal performance. Considering aforementioned challenges, we propose Hypergraph-based Multi-instance Contrastive Reinforcement learning model (HeMiCoRe). HeMiCoRe employs novel...
Abstract To develop and validate a nomogram for predicting the risk of adverse events (intraoperative massive haemorrhage or retained products conception) associated with termination Caesarean scar pregnancy (CSP). Data were retrospectively collected from patients diagnosed CSP who underwent Dilation Curettage (D&C) at two hospitals. This data was divided into internal external cohorts analysis. The cohort randomly split, 70% designated training set 30% an validation set. served...
<title>Abstract</title> Purpose To construct and validate a nomogram to predict the risk of adverse events (intraoperative massive hemorrhage or retained products conception) during termination Cesarean scar pregnancy (CSP). Method Data from patients diagnosed with CSP who underwent Dilation Curettage (D&C) at two hospitals were retrospectively collected. This data formed both internal external cohorts for analysis. The cohort was split randomly, 70% allocated training set 30% an...
In this study, a deep learning model based on quantum chemistry is introduced to enhance the accuracy and efficiency of predicting DNA reaction parameters. By integrating chemical calculations with self-designed descriptor matrices, offers comprehensive description energy variations considers broad range relevant factors. To overcome challenge limited labeled data, an active method employed. The results demonstrate that outperforms existing methods in hybridization free energies strand...
Purpose: To determine stereotactic body radiation therapy (SBRT) dose tolerance limits and to develop a way conveniently ensure the are being met for each patient. Methods Materials: An extensive literature review was conducted emerging critical structures SBRT treatments of various cranial extracranial targets. We developed program called DVH Evaluator overlay appropriate onto volume histogram (DVH) results, highlight any instances where limit goals not met. This can be used interactively...
Many classification models have been proposed in past few decades. Lots of variations based on those are also developed for better performance. Instead model tuning or modification, we achieve higher accuracy by analyzing the dataset and recovering instances that mis-classified given classifier. We develop three metrics to identify instances. Experiments show our method can obtain performance improvement with chosen classifier multiple datasets.