Ni Wang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Ni Wang has a total of 51 co-authors. Here are the top co-authors by number of shared publications:
Alaa Moamer
7 shared publications- Gang Yan
School of Artificial Intelligence, Hebei University of Technology, Tianjin, China
7 shared publications - Girija Daliah
Department of Medicine, Cancer Research Program, McGill University Health Centre, Montreal, Canada
7 shared publications - Honglei Liu
Capital Medical University
7 shared publications - Hui Chen
School of Biomedical Engineering, Capital Medical University, Beijing, China
7 shared publications - Jean-Jacques Lebrun
Department of Medicine, Cancer Research Program, McGill University Health Centre, Montreal, Canada
7 shared publications
Latest Publications
- Open Access
- 24 July 2026
- Published by Elsevier in SSRN Electronic Journal
Mathematically Guided Deep Denoising Enables Fast Scanning Electron Microscopy Characterization of Material Microstructures with High Fidelity
Fast image acquisition in scanning electron microscopy is essential for materials characterization at high throughput, but reduced electron beam dwell time introduces stochastic noise that obscures microstructural interfaces, precipitates, fracture features, and particle boundaries. To address the conflict between rapid scanning and structural preservation, this study proposes a deep learning-based adaptive denoising network for scanning electron microscope image denoising. The method integrates dilated reparameterized convolution blocks and dynamic convolution blocks within an encoder-decoder architecture to balance global contextual modeling and local feature preservation. Unlike conventional approaches relying on empirical noise assumptions, the proposed method is trained on a synthetic dataset constructed from mathematically derived noise priors for scanning electron microscope imaging. Quantitative evaluation shows that the proposed method outperforms the compared methods under diverse noise regimes. In the two most severe simulated scanning conditions, it achieves peak signal-to-noise ratio values of 35.94 decibels and 40.62 decibels, structural similarity index values of 0.9344 and 0.9613, and learned perceptual image patch similarity values of 0.0926 and 0.0649. In addition, the method shows promising generalization to real-world industrial and scientific scanning electron microscope images, supporting high-throughput materials characterization while preserving image quality. The code and curated test set will be available at our project repository.
- 24 July 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMathematically Guided Deep Denoising Enables Fast Scanning Electron Microscopy Characterization of Material Microstructures with High Fidelity
Fast image acquisition in scanning electron microscopy is essential for materials characterization at high throughput, but reduced electron beam dwell time introduces stochastic noise that obscures microstructural interfaces, precipitates, fracture features, and particle boundaries. To address the conflict between rapid scanning and structural preservation, this study proposes a deep learning-based adaptive denoising network for scanning electron microscope image denoising. The method integrates dilated reparameterized convolution blocks and dynamic convolution blocks within an encoder-decoder architecture to balance global contextual modeling and local feature preservation. Unlike conventional approaches relying on empirical noise assumptions, the proposed method is trained on a synthetic dataset constructed from mathematically derived noise priors for scanning electron microscope imaging. Quantitative evaluation shows that the proposed method outperforms the compared methods under diverse noise regimes. In the two most severe simulated scanning conditions, it achieves peak signal-to-noise ratio values of 35.94 decibels and 40.62 decibels, structural similarity index values of 0.9344 and 0.9613, and learned perceptual image patch similarity values of 0.0926 and 0.0649. In addition, the method shows promising generalization to real-world industrial and scientific scanning electron microscope images, supporting high-throughput materials characterization while preserving image quality. The code and curated test set will be available at our project repository.
- Open Access
- 31 March 2023
- Published by American Association for Cancer Research (AACR)
Data from Differential Regulation of Cancer Progression by CDK4/6 Plays a Central Role in DNA Replication and Repair Pathways
Although the cyclin-dependent kinases CDK4 and CDK6 play fundamental roles in cancer, the specific pathways and downstream targets by which they exert their tumorigenic effects remain elusive. In this study, we uncover distinct and novel functions for these kinases in regulating tumor formation and metastatic colonization in various solid tumors, including those of the breast, prostate, and pancreas. Combining in vivo CRISPR-based CDK4 and CDK6 gene editing with pharmacologic inhibition approaches in orthotopic transplantation and patient-derived xenograft preclinical models, we defined clear functions for CDK4 and CDK6 in facilitating tumor growth and progression in metastatic cancers. Transcriptomic profiling of CDK4/6 CRISPR knockouts in breast cancer revealed these two kinases to regulate cancer progression through distinct mechanisms. CDK4 regulated prometastatic inflammatory cytokine signaling, whereas CDK6 mainly controlled DNA replication and repair processes. Inhibition of CDK6 but not CDK4 resulted in defective DNA repair and increased DNA damage. Multiple CDK6 DNA replication/repair genes were not only associated with cancer subtype, grades, and poor clinical outcomes, but also facilitated primary tumor growth and metastasis in vivo. CRISPR-based genomic deletion of CDK6 efficiently blocked tumor formation and progression in preestablished cell- and patient-derived xenograft preclinical models of breast cancer, providing a potential novel targeted therapy for these deadly tumors.Significance:In-depth transcriptomic analysis identifies cyclin-dependent kinases CDK4 and CDK6 as regulators of metastasis through distinct signaling pathways and reveals the DNA replication/repair pathway as central in promoting these effects.
- 31 March 2023
- Published by American Association for Cancer Research (AACR)
Open AccessData from Differential Regulation of Cancer Progression by CDK4/6 Plays a Central Role in DNA Replication and Repair Pathways
Although the cyclin-dependent kinases CDK4 and CDK6 play fundamental roles in cancer, the specific pathways and downstream targets by which they exert their tumorigenic effects remain elusive. In this study, we uncover distinct and novel functions for these kinases in regulating tumor formation and metastatic colonization in various solid tumors, including those of the breast, prostate, and pancreas. Combining in vivo CRISPR-based CDK4 and CDK6 gene editing with pharmacologic inhibition approaches in orthotopic transplantation and patient-derived xenograft preclinical models, we defined clear functions for CDK4 and CDK6 in facilitating tumor growth and progression in metastatic cancers. Transcriptomic profiling of CDK4/6 CRISPR knockouts in breast cancer revealed these two kinases to regulate cancer progression through distinct mechanisms. CDK4 regulated prometastatic inflammatory cytokine signaling, whereas CDK6 mainly controlled DNA replication and repair processes. Inhibition of CDK6 but not CDK4 resulted in defective DNA repair and increased DNA damage. Multiple CDK6 DNA replication/repair genes were not only associated with cancer subtype, grades, and poor clinical outcomes, but also facilitated primary tumor growth and metastasis in vivo. CRISPR-based genomic deletion of CDK6 efficiently blocked tumor formation and progression in preestablished cell- and patient-derived xenograft preclinical models of breast cancer, providing a potential novel targeted therapy for these deadly tumors.Significance:In-depth transcriptomic analysis identifies cyclin-dependent kinases CDK4 and CDK6 as regulators of metastasis through distinct signaling pathways and reveals the DNA replication/repair pathway as central in promoting these effects.
- Open Access
- 6 January 2022
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 24 (1) , e30720
Sequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development
Background: Sequential information in electronic medical records is valuable and helpful for patient outcome prediction but is rarely used for patient similarity measurement because of its unevenness, irregularity, and heterogeneity. Objective: We aimed to develop a patient similarity framework for patient outcome prediction that makes use of sequential and cross-sectional information in electronic medical record systems. Methods: Sequence similarity was calculated from timestamped event sequences using edit distance, and trend similarity was calculated from time series using dynamic time warping and Haar decomposition. We also extracted cross-sectional information, namely, demographic, laboratory test, and radiological report data, for additional similarity calculations. We validated the effectiveness of the framework by constructing k–nearest neighbors classifiers to predict mortality and readmission for acute myocardial infarction patients, using data from (1) a public data set and (2) a private data set, at 3 time points—at admission, on Day 7, and at discharge—to provide early warning patient outcomes. We also constructed state-of-the-art Euclidean-distance k–nearest neighbor, logistic regression, random forest, long short-term memory network, and recurrent neural network models, which were used for comparison. Results: With all available information during a hospitalization episode, predictive models using the similarity model outperformed baseline models based on both public and private data sets. For mortality predictions, all models except for the logistic regression model showed improved performances over time. There were no such increasing trends in predictive performances for readmission predictions. The random forest and logistic regression models performed best for mortality and readmission predictions, respectively, when using information from the first week after admission. Conclusions: For patient outcome predictions, the patient similarity framework facilitated sequential similarity calculations for uneven electronic medical record data and helped improve predictive performance.
- 6 January 2022
- Published by JMIR Publications Inc. in Journal of Medical Internet Research
- Vol. 24 (1) , e30720
Open AccessSequential Data–Based Patient Similarity Framework for Patient Outcome Prediction: Algorithm Development
Background: Sequential information in electronic medical records is valuable and helpful for patient outcome prediction but is rarely used for patient similarity measurement because of its unevenness, irregularity, and heterogeneity. Objective: We aimed to develop a patient similarity framework for patient outcome prediction that makes use of sequential and cross-sectional information in electronic medical record systems. Methods: Sequence similarity was calculated from timestamped event sequences using edit distance, and trend similarity was calculated from time series using dynamic time warping and Haar decomposition. We also extracted cross-sectional information, namely, demographic, laboratory test, and radiological report data, for additional similarity calculations. We validated the effectiveness of the framework by constructing k–nearest neighbors classifiers to predict mortality and readmission for acute myocardial infarction patients, using data from (1) a public data set and (2) a private data set, at 3 time points—at admission, on Day 7, and at discharge—to provide early warning patient outcomes. We also constructed state-of-the-art Euclidean-distance k–nearest neighbor, logistic regression, random forest, long short-term memory network, and recurrent neural network models, which were used for comparison. Results: With all available information during a hospitalization episode, predictive models using the similarity model outperformed baseline models based on both public and private data sets. For mortality predictions, all models except for the logistic regression model showed improved performances over time. There were no such increasing trends in predictive performances for readmission predictions. The random forest and logistic regression models performed best for mortality and readmission predictions, respectively, when using information from the first week after admission. Conclusions: For patient outcome predictions, the patient similarity framework facilitated sequential similarity calculations for uneven electronic medical record data and helped improve predictive performance.