Yanjun Wu

ORCID: 0000-0003-0233-2917
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Image Retrieval and Classification Techniques
  • Forensic and Genetic Research
  • Forensic Anthropology and Bioarchaeology Studies
  • Advanced Image and Video Retrieval Techniques
  • Image Processing and 3D Reconstruction
  • Advanced X-ray and CT Imaging
  • Domain Adaptation and Few-Shot Learning
  • Radiomics and Machine Learning in Medical Imaging
  • Phytochemicals and Medicinal Plants
  • Adversarial Robustness in Machine Learning
  • Interstitial Lung Diseases and Idiopathic Pulmonary Fibrosis
  • Aerogels and thermal insulation
  • Traditional Chinese Medicine Analysis
  • Heavy Metal Exposure and Toxicity
  • Advanced Nanomaterials in Catalysis
  • Advanced Computational Techniques and Applications
  • Multimodal Machine Learning Applications
  • Ferroptosis and cancer prognosis
  • S100 Proteins and Annexins
  • Artificial Intelligence in Healthcare and Education
  • Cell Image Analysis Techniques
  • Diagnosis and treatment of tuberculosis
  • Medical Imaging and Analysis
  • Handwritten Text Recognition Techniques
  • Advanced Neural Network Applications

Chengde Medical University
2022-2025

Southern Medical University
2023-2024

Guangzhou Center for Disease Control and Prevention
2024

Shanghai Institute of Nutrition and Health
2018-2024

University of Chinese Academy of Sciences
2018-2024

Guangdong Province Women and Children Hospital
2023

South China University of Technology
2021

Dalian Maritime University
2016-2019

Shanghai Institutes for Biological Sciences
2018

GlaxoSmithKline (United Kingdom)
2007-2010

Abstract Background Radiomic diagnosis models generally consider only a single dimension of information, leading to limitations in their diagnostic accuracy and reliability. The integration multiple dimensions information into the deep learning model have potential improve its capabilities. purpose study was evaluate performance distinguishing tuberculosis (TB) nodules lung cancer (LC) based on features, radiomic clinical information. Methods Positron emission tomography (PET) computed (CT)...

10.1111/1759-7714.14924 article EN cc-by-nc Thoracic Cancer 2023-05-14

Pancreatic cancer is regarded as the most fatal and aggressive malignancy due to its low 5‐year survival rate poor prognosis. The approaches of early diagnosis treatment are limited, which makes it urgent identify complex mechanism pancreatic oncogenesis. In this study, we used RNA‐seq investigate transcriptomic (mRNA miRNA) profiles in paired tumor normal samples from ten patients. More than 1000 differentially expressed genes were identified, nearly half also found be majority examined...

10.1002/mc.22920 article EN Molecular Carcinogenesis 2018-10-08

Abstract Background This study investigated whether the Combat compensation method can remove variability of radiomic features extracted from different scanners, while also examining its impact on subsequent predictive performance machine learning models. Materials and methods 135 CT images Credence Cartridge Radiomic phantoms were collected screened three scanners manufactured by Siemens, Philips, GE. 100 20 according to Lasso regression method. The rubber resin-filled regions in cartridges...

10.1186/s12880-024-01306-4 article EN cc-by BMC Medical Imaging 2024-06-06

Given a query shoeprint image, retrieval aims to retrieve the most similar shoeprints available from large set of images. Most existing approaches focus on designing single low-level features highlight aspects shoeprints, but their precision may vary dramatically with quality and content Therefore, in this paper, we proposed method enhance two perspectives: (i) integrate strengths three kinds yield more satisfactory results; (ii) traditional distance-based similarity by leveraging...

10.3390/info10020045 article EN cc-by Information 2019-01-30

Isobaric labeling-based mass spectrometry (ILMS) has been widely used to quantify, on a proteome-wide scale, the relative protein abundance in different biological conditions. However, large-scale ILMS data sets typically involve multiple runs of spectrometry, bringing great computational difficulty integration samples. We present zMAP, toolset that makes intensities comparable across by modeling associated mean-variance dependence and accordingly applying variance stabilizing...

10.1186/s13059-024-03382-9 article EN cc-by-nc-nd Genome biology 2024-10-14

Lead (Pb) is a pervasive toxic metal contaminant associated with high risk of myocardial injury. However, the precise mechanism underlying Pb-induced injury has yet to be fully elucidated. In this study, murine model Pb exposure (0, 1, 5, and 10 mg/kg) was employed investigate involvement neutrophil degranulation in induction Notably, serum levels cardiac troponin I (cTnI) creatine kinase-MB (CK-MB) increased significantly Pb-exposed mice, whereas cTnI cardiomyocytes decreased, suggesting...

10.1016/j.ecoenv.2023.115694 article EN cc-by-nc-nd Ecotoxicology and Environmental Safety 2023-11-18

As a kind of forensic evidence, shoeprint conveys many important human characteristics, and it plays vital role in investigations. Millions shoeprints are acquired from crime scenes, is challenging task to retrieve the most similar for query shoeprint. Most retrieval methods sort images by feature similarities with respect shoeprint; however, results not always what investigator expects because algorithm cannot determine prefers based on only content This paper proposes method guide process...

10.1109/access.2019.2912585 article EN cc-by-nc-nd IEEE Access 2019-01-01

As a kind of forensic evidence, shoeprints have been treated as important fingerprint and DNA evidence in investigations. Shoeprint verification is used to determine whether two could, or could not, made by the same shoe. Successful shoeprint has tremendous evidentiary value, result can link suspect crime, even crime scenes each other. In practice, manually performed experts; however, it too dependent on experts' experience. This meaningful challenging problem, there are few attempts tackle...

10.3390/s19112491 article EN cc-by Sensors 2019-05-31

10.7544/issn1000-1239.2018.20170252 article EN Journal of Computer Research and Development 2018-07-01

The development of novel methodologies that utilize various non-invasive imaging modalities has resulted in their increased use and relevance the pie-clinical phase pharmaceutical compound development. Scientific questions may benefit from techniques are often more robustly addressed with data acquired complementary modalities, intensifying need for performing cross-modality image registration fusion. In this work, a methodology multimodal coregistration fusion MRI PET volumes is presented...

10.1109/mmsp.2007.4412883 article EN 2007-01-01

We reported a multi-level composite SERS platform which could provide new method for accurate and quantitative trace detection pharmacodynamic substance in traditional Chinese medicine (TCM). Using natural dragonfly wings as the template, Ag NPs were deposited on V-Ti nano-columns through controllable magnetron sputtering technology. Ag30@V-Ti20@D.W possessed uniform nanostructure exhibited High-performance signal sensing capability. The limit (LOD) of R6G was 1×10-10 M. Through machine...

10.2139/ssrn.4463821 preprint EN 2023-01-01

10.7544/issn1000-1239.2017.20170177 article EN Journal of Computer Research and Development 2017-08-01

Lead (Pb) is a toxic metal contaminant that can increase the risk of myocardial injury. However, mechanism injury induction by Pb remains to be elucidated. Herein, mice model exposure (0, 1, 5, and 10 mg/kg) was established understand role neutrophil degranulation in Pb-induced Levels cardiac troponin I (cTnI) creatine kinase-MB (CK-MB) serum exposed significantly increased but cTnI levels cardiomyocytes decreased. Thus, induce early Furthermore, promoted as observed myeloperoxidase (MPO)...

10.2139/ssrn.4293077 article EN SSRN Electronic Journal 2022-01-01

Giant molecules self-assemble into long-range ordered lamellar structures in bulk with ultra-small domain sizes (down to 3 nm), driven by the strong tendency and fast kinetics of crystallization for vinyl polyhedral oligomeric silsesquioxane (VPOSS) segment. This crystallization-induced self-assembly strategy may offer an opportunity create integrated circuits (ICs) feature lithographic techniques microelectronic industry. More information can be found Communication Z.-H. Guo et al. (DOI:...

10.1002/chem.202101552 article EN Chemistry - A European Journal 2021-05-14
Coming Soon ...