Kaiyuan Yang
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
Kaiyuan Yang has a total of 159 co-authors. Here are the top co-authors by number of shared publications:
- Andrew Makmur
Department of Diagnostic Imaging, National University Hospital, Singapore, Singapore
2 shared publications - Bjoern Menze
Department of Quantitative Biomedicine, University of Zurich, Zurich, Switzerland
2 shared publications - Daniele Falcetta
EURECOM
2 shared publications - James Thomas Patrick Decourcy Hallinan
Department of Diagnostic Imaging, National University Hospital, Singapore, Singapore
2 shared publications - Maria Alejandra Zuluaga
EURECOM
2 shared publications Olafs Vandans
Synspective Inc
2 shared publications
Latest Publications
- 23 July 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (8)
The TopCoW Challenge — Topology-Aware Circle of Willis Segmentation for CT and MR Angiography
- 23 July 2026
- Published by Massachusetts Medical Society in NEJM AI
- Vol. 3 (8)
The TopCoW Challenge — Topology-Aware Circle of Willis Segmentation for CT and MR Angiography
- 20 September 2025
- Published by Springer Nature
- p. 655-665
VesselVerse: A Dataset and Collaborative Framework for Vessel Annotation
- 20 September 2025
- Published by Springer Nature
- p. 655-665
VesselVerse: A Dataset and Collaborative Framework for Vessel Annotation
- Top Cited Papers
- 1 July 2021
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 300 (1) , 130-138
Deep Learning Model for Automated Detection and Classification of Central Canal, Lateral Recess, and Neural Foraminal Stenosis at Lumbar Spine MRI
A deep learning model detected and classified lumbar spinal stenosis at MRI comparable with radiologists for central canal and lateral recess stenosis, with slightly lower agreement for neural fora... Background Assessment of lumbar spinal stenosis at MRI is repetitive and time consuming. Deep learning (DL) could improve productivity and the consistency of reporting. Purpose To develop a DL mod...
- 1 July 2021
- Published by Radiological Society of North America (RSNA) in Radiology
- Vol. 300 (1) , 130-138
Top Cited PapersDeep Learning Model for Automated Detection and Classification of Central Canal, Lateral Recess, and Neural Foraminal Stenosis at Lumbar Spine MRI
A deep learning model detected and classified lumbar spinal stenosis at MRI comparable with radiologists for central canal and lateral recess stenosis, with slightly lower agreement for neural fora... Background Assessment of lumbar spinal stenosis at MRI is repetitive and time consuming. Deep learning (DL) could improve productivity and the consistency of reporting. Purpose To develop a DL mod...
- 1 April 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1718-1721
ForkBase: Immutable, Tamper-evident Storage Substrate for Branchable Applications
Data collaboration activities typically require systematic or protocol-based coordination to be scalable. Git, an effective enabler for collaborative coding, has been attested for its success in countless projects around the world. Hence, applying the Git philosophy to general data collaboration beyond coding is motivating. We call it Git for data. However, the original Git design handles data at the file granule, which is considered too coarse-grained for many database applications. We argue that Git for data should be co-designed with database systems. To this end, we developed ForkBase to make Git for data practical. ForkBase is a distributed, immutable storage system designed for data version management and data collaborative operation. In this demonstration, we show how ForkBase can greatly facilitate collaborative data management and how its novel data deduplication technique can improve storage efficiency for archiving massive data versions.
- 1 April 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1718-1721
ForkBase: Immutable, Tamper-evident Storage Substrate for Branchable Applications
Data collaboration activities typically require systematic or protocol-based coordination to be scalable. Git, an effective enabler for collaborative coding, has been attested for its success in countless projects around the world. Hence, applying the Git philosophy to general data collaboration beyond coding is motivating. We call it Git for data. However, the original Git design handles data at the file granule, which is considered too coarse-grained for many database applications. We argue that Git for data should be co-designed with database systems. To this end, we developed ForkBase to make Git for data practical. ForkBase is a distributed, immutable storage system designed for data version management and data collaborative operation. In this demonstration, we show how ForkBase can greatly facilitate collaborative data management and how its novel data deduplication technique can improve storage efficiency for archiving massive data versions.
- 11 February 2020
- Published by American Physical Society (APS) in Physical Review E
- Vol. 101 (2) , 022502
Identifying knot types of polymer conformations by machine learning
We investigate the use of artificial neural networks (NNs) as an alternative tool to current analytical methods for recognizing knots in a given polymer conformation. The motivation is twofold. First, it is of interest to examine whether NNs are effective at learning the global and sequential properties that uniquely define a knot. Second, knot classification is an important and unsolved problem in mathematical and physical sciences, and NNs may provide insights into this problem. Motivated by these points, we generate millions of polymer conformations for five knot types: 0, , and , and we design various NN models for classification. Our best model achieves a five-class classification accuracy of above on a polymer of 100 monomers. We find that the sequential modeling ability of recurrent NNs is crucial for this result, as it outperforms feed-forward NNs and successfully generalizes to differently sized conformations as well. We present our methods and suggest that deep learning may be used in specific applications of knot detection where some error is permissible. Hopefully, with further development, NNs can offer an alternative computational method for knot identification and facilitate knot research in mathematical and physical sciences.- 11 February 2020
- Published by American Physical Society (APS) in Physical Review E
- Vol. 101 (2) , 022502
Identifying knot types of polymer conformations by machine learning
We investigate the use of artificial neural networks (NNs) as an alternative tool to current analytical methods for recognizing knots in a given polymer conformation. The motivation is twofold. First, it is of interest to examine whether NNs are effective at learning the global and sequential properties that uniquely define a knot. Second, knot classification is an important and unsolved problem in mathematical and physical sciences, and NNs may provide insights into this problem. Motivated by these points, we generate millions of polymer conformations for five knot types: 0, , and , and we design various NN models for classification. Our best model achieves a five-class classification accuracy of above on a polymer of 100 monomers. We find that the sequential modeling ability of recurrent NNs is crucial for this result, as it outperforms feed-forward NNs and successfully generalizes to differently sized conformations as well. We present our methods and suggest that deep learning may be used in specific applications of knot detection where some error is permissible. Hopefully, with further development, NNs can offer an alternative computational method for knot identification and facilitate knot research in mathematical and physical sciences.- Open Access
- 16 November 2017
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 12 (11) , e0188015
From bead to rod: Comparison of theories by measuring translational drag coefficients of micron-sized magnetic bead-chains in Stokes flow
Frictional drag force on an object in Stokes flow follows a linear relationship with the velocity of translation and a translational drag coefficient. This drag coefficient is related to the size, shape, and orientation of the object. For rod-like objects, analytical solutions of the drag coefficients have been proposed based on three rough approximations of the rod geometry, namely the bead model, ellipsoid model, and cylinder model. These theories all agree that translational drag coefficients of rod-like objects are functions of the rod length and aspect ratio, but differ among one another on the correction factor terms in the equations. By tracking the displacement of the particles through stationary fluids of calibrated viscosity in magnetic tweezers setup, we experimentally measured the drag coefficients of micron-sized beads and their bead-chain formations with chain length of 2 to 27. We verified our methodology with analytical solutions of dimers of two touching beads, and compared our measured drag coefficient values of rod-like objects with theoretical calculations. Our comparison reveals several analytical solutions that used more appropriate approximation and derived formulae that agree with our measurement better.
- 16 November 2017
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 12 (11) , e0188015
Open AccessFrom bead to rod: Comparison of theories by measuring translational drag coefficients of micron-sized magnetic bead-chains in Stokes flow
Frictional drag force on an object in Stokes flow follows a linear relationship with the velocity of translation and a translational drag coefficient. This drag coefficient is related to the size, shape, and orientation of the object. For rod-like objects, analytical solutions of the drag coefficients have been proposed based on three rough approximations of the rod geometry, namely the bead model, ellipsoid model, and cylinder model. These theories all agree that translational drag coefficients of rod-like objects are functions of the rod length and aspect ratio, but differ among one another on the correction factor terms in the equations. By tracking the displacement of the particles through stationary fluids of calibrated viscosity in magnetic tweezers setup, we experimentally measured the drag coefficients of micron-sized beads and their bead-chain formations with chain length of 2 to 27. We verified our methodology with analytical solutions of dimers of two touching beads, and compared our measured drag coefficient values of rod-like objects with theoretical calculations. Our comparison reveals several analytical solutions that used more appropriate approximation and derived formulae that agree with our measurement better.
- Open Access
- 6 July 2017
- Published by Springer Nature in Scientific Reports
- Vol. 7 (1) , 4783
Mechanistic action of weak acid drugs on biofilms
Selective permeability of a biofilm matrix to some drugs has resulted in the development of drug tolerant bacteria. Here we studied the efficacy of a weak organic acid drug, N-acetyl-L-cysteine (NAC), on the eradication of biofilms formed by the mucoid strain of Pseudomonas aeruginosa and investigated the commonality of this drug with that of acetic acid. We showed that NAC and acetic acid at pH < pKa can penetrate the matrix and eventually kill 100% of the bacteria embedded in the biofilm. Once the bacteria are killed, the microcolonies swell in size and passively shed bacteria, suggesting that the bacteria act as crosslinkers within the extracellular matrix. Despite shedding of the bacteria, the remnant matrix remains intact and behaves as a pH-responsive hydrogel. These studies not only have implications for drug design but also offer a route to generate robust soft matter materials.
- 6 July 2017
- Published by Springer Nature in Scientific Reports
- Vol. 7 (1) , 4783
Open AccessMechanistic action of weak acid drugs on biofilms
Selective permeability of a biofilm matrix to some drugs has resulted in the development of drug tolerant bacteria. Here we studied the efficacy of a weak organic acid drug, N-acetyl-L-cysteine (NAC), on the eradication of biofilms formed by the mucoid strain of Pseudomonas aeruginosa and investigated the commonality of this drug with that of acetic acid. We showed that NAC and acetic acid at pH < pKa can penetrate the matrix and eventually kill 100% of the bacteria embedded in the biofilm. Once the bacteria are killed, the microcolonies swell in size and passively shed bacteria, suggesting that the bacteria act as crosslinkers within the extracellular matrix. Despite shedding of the bacteria, the remnant matrix remains intact and behaves as a pH-responsive hydrogel. These studies not only have implications for drug design but also offer a route to generate robust soft matter materials.
- Open Access
- 14 September 2012
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 7 (9) , e44988
High Glucose Predisposes Gene Expression and ERK Phosphorylation to Apoptosis and Impaired Glucose-Stimulated Insulin Secretion via the Cytoskeleton
Chronic high glucose (HG) inflicts glucotoxicity on vulnerable cell types such as pancreatic β cells and contributes to insulin resistance and impaired insulin secretion in diabetic patients. To identify HG-induced cellular aberrations that are candidate mediators of glucotoxicity in pancreatic β cells, we analyzed gene expression in ERoSHK6, a mouse insulin-secreting cell line after chronic HG exposure (six-day exposure to 33.3 mM glucose). Chronic HG exposure which reduced glucose-stimulated insulin secretion (GSIS) increased transcript levels of 185 genes that clustered primarily in 5 processes namely cellular growth and proliferation; cell death; cellular assembly and organization; cell morphology; and cell-to-cell signaling and interaction. The former two were validated by increased apoptosis of ERoSHK6 cells after chronic HG exposure and reaffirmed the vulnerability of β cells to glucotoxicity. The three remaining processes were partially substantiated by changes in cellular morphology and structure, and instigated an investigation of the cytoskeleton and cell-cell adhesion. These studies revealed a depolymerized actin cytoskeleton that lacked actin stress fibers anchored at vinculin-containing focal adhesion sites as well as loss of E-cadherin-mediated cell-cell adherence after exposure to chronic HG, and were concomitant with constitutive ERK1/2 phosphorylation that was refractory to serum and glucose deprivation. Although inhibition of ERK phosphorylation by PD98059 promoted actin polymerization, it increased apoptosis and GSIS impairment. These findings suggest that ERK phosphorylation is a proximate regulator of cellular processes targeted by chronic HG-induced gene expression and that dynamic actin polymerization and depolymerization is important in β cell survival and function. Therefore, chronic HG alters gene expression and signal transduction to predispose the cytoskeleton towards apoptosis and GSIS impairment.
- 14 September 2012
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 7 (9) , e44988
Open AccessHigh Glucose Predisposes Gene Expression and ERK Phosphorylation to Apoptosis and Impaired Glucose-Stimulated Insulin Secretion via the Cytoskeleton
Chronic high glucose (HG) inflicts glucotoxicity on vulnerable cell types such as pancreatic β cells and contributes to insulin resistance and impaired insulin secretion in diabetic patients. To identify HG-induced cellular aberrations that are candidate mediators of glucotoxicity in pancreatic β cells, we analyzed gene expression in ERoSHK6, a mouse insulin-secreting cell line after chronic HG exposure (six-day exposure to 33.3 mM glucose). Chronic HG exposure which reduced glucose-stimulated insulin secretion (GSIS) increased transcript levels of 185 genes that clustered primarily in 5 processes namely cellular growth and proliferation; cell death; cellular assembly and organization; cell morphology; and cell-to-cell signaling and interaction. The former two were validated by increased apoptosis of ERoSHK6 cells after chronic HG exposure and reaffirmed the vulnerability of β cells to glucotoxicity. The three remaining processes were partially substantiated by changes in cellular morphology and structure, and instigated an investigation of the cytoskeleton and cell-cell adhesion. These studies revealed a depolymerized actin cytoskeleton that lacked actin stress fibers anchored at vinculin-containing focal adhesion sites as well as loss of E-cadherin-mediated cell-cell adherence after exposure to chronic HG, and were concomitant with constitutive ERK1/2 phosphorylation that was refractory to serum and glucose deprivation. Although inhibition of ERK phosphorylation by PD98059 promoted actin polymerization, it increased apoptosis and GSIS impairment. These findings suggest that ERK phosphorylation is a proximate regulator of cellular processes targeted by chronic HG-induced gene expression and that dynamic actin polymerization and depolymerization is important in β cell survival and function. Therefore, chronic HG alters gene expression and signal transduction to predispose the cytoskeleton towards apoptosis and GSIS impairment.