Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks132 from public data
- BANKSY unifies cell typing and tissue domain segmentation for scalable spatial omics data analysis273
BANKSY is an algorithm with R and Python implementations that identifies both cell types and tissue domains from spatially resolved omics data by incorporating spatial kernels capturing microenvironmental information and is scalable to millions of cells.
- Tumor‐adjacent tissue co‐expression profile analysis reveals pro‐oncogenic ribosomal gene signature for prognosis of resectable hepatocellular carcinoma236
The common co‐transcriptional pattern of ribosome biogenesis genes in PT and AT from HCC patients suggests a new scalable prognostic system, as supported by the model of tumor‐like metabolic redirection/assimilation in non‐malignant AT.
- Fence GAN: Towards Better Anomaly Detection111
With the modified GAN loss proposed, the anomaly detection method, called Fence GAN (FGAN), directly uses the discriminator score as an anomaly threshold and the experimental results show that FGAN yields the best anomaly classification accuracy compared to state-of-the-art methods.
- Anterior Segment Optical Coherence Tomography Parameters in Subtypes of Primary Angle Closure99
Eyes with APAC had the narrowest angles, smallest anterior segment dimensions, thickest iris, and largest LV compared with PACS, PAC, and PACG.
- Convergence and refinement of the Wang–Landau algorithm74
To understand the mechanism behind the Wang–Landau method, an error analysis found that a steady state is reached where the fluctuations in the accumulated energy histogram saturate at values proportional to [ log ( f ) ] − 1 / 2 .
- Large‐scale image‐based screening and profiling of cellular phenotypes73
The focus is on phenotypic profiling, a computational procedure for constructing quantitative and compact representations of cellular phenotypes based on the images collected in these screens.
- Quantitative neurite outgrowth measurement based on image segmentation with topological dependence73
It is shown that the transfection of Toca‐1 cDNA induces longer neurites with more complexities than serum starvation, and a tracing algorithm was developed to automatically trace neurites and measure their lengths quantitatively on a cell‐by‐cell basis.
- Gated-Dilated Networks for Lung Nodule Classification in CT Scans71
Compared to the baseline models, the proposed Gated-Dilated networks improves the classification accuracies of mid-range sized nodules and observes a relationship between the size of the nodule and the attention signal generated by the Context-Aware sub-network, which validates the new network architecture.
- Myopia in Asian Subjects with Primary Angle Closure63
With the increasing rate of myopia in many East Asian populations, there may be many subjects with axial myopia but shallow ACD and angle closure, and ophthalmologists should not assume that glaucoma patients who are myopic have open angles.
- Subgrouping of Primary Angle-Closure Suspects Based on Anterior Segment Optical Coherence Tomography Parameters62
Clustering analysis identified 3 distinct subgroups of PACS subjects based on anterior segment optical coherence tomography and biometric parameters that may be relevant for understanding angle-closure pathogenesis and management.
- An end-to-end breast tumour classification model using context-based patch modelling – A BiLSTM approach for image classification56
This work found out that BiLSTMs with CNN features have performed much better in modelling patches into an end-to-end Image classification network and the variable dimensions of WSI tumour regions were used for classification without the need for resizing, suggesting that the method is independent of tumour image size and can process large dimensional images without losing the resolution details.
- Mapping the Monte Carlo Scheme to Langevin Dynamics: A Fokker-Planck Approach55
This work derives the drift and diffusion FPE terms corresponding to the MC method and shows that it is analytically equivalent to the stochastic Landau-Lifshitz-Gilbert (LLG) equation of Langevin-based micromagnetics.
- Evolving generalized Voronoi diagrams for accurate cellular image segmentation52
Utilizing image intensity and geometric information, the evolving generalized Voronoi diagram (EGVD) algorithm preserves topological dependence easily in both 2D and 3D images, such that touching cells can be segmented satisfactorily.
- Automated grading of acne vulgaris by deep learning with convolutional neural networks48
The visual assessment and severity grading of acne vulgaris by physicians can be subjective, resulting in inter‐ and intra‐observer variability.
- 47
- How bioinformatics influences health informatics: usage of biomolecular sequences, expression profiles and automated microscopic image analyses for clinical needs and public health45
Advances in the area of cancer biomarker discovery, in the clinically relevant characterization of patient-specific viral and bacterial pathogens, as well as progress in the automated assessment of histopathological images are exemplarily reviewed.
- Bioimage informatics approach to automated meibomian gland analysis in infrared images of meibography43
A method of automated gland segmentation which allows images to be classified into healthy and unhealthy classes is developed and achieved a 100% accuracy, but 7/38 intermediate images were incorrectly classified.
- Automated Renal Cancer Grading Using Nuclear Pleomorphic Patterns36
An automated, image-based system that classifies ccRCC slides by quantifying nuclear pleomorphic patterns in an objective and consistent interpretable fashion can aid pathologists in histopathologic assessment and can be extended to other cancers whose corresponding grading systems use nuclear pattern information.
- Resolution enhancement and realistic speckle recovery with generative adversarial modeling of micro-optical coherence tomography35
This preliminary study suggests that deep learning generative models trained on OCT images from high-performance prototype systems may have potential in enhancing lower resolution data from mainstream/commercial systems, thereby bringing cutting-edge technology to the masses at low cost.
- Nuclear import of a lipid-modified transcription factor35
NFAT5a is an example of TFs immobilized with lipid anchors at cyotoplasmic membranes in the resting state and that, nevertheless, can translocate into the nucleus upon signal induction.
- A comprehensive AI model development framework for consistent Gleason grading32
An AI-based system which automatically checks image quality, standardizes the appearance of images from different equipment, learns from pathologists’ feedback, and constantly improves model performance is built, which could potentially improve prostate cancer diagnosis and management.
- Gland segmentation in prostate histopathological images32
Experimental results show that the proposed automated gland segmentation system has good performance and can be a promising tool to help decrease interobserver variability among pathologists.
- Light sheet fluorescence microscopy (LSFM): past, present and future28
The aim of this paper is to facilitate the set-up and use of LSFM by reviewing and comparing open access projects, image processing tools and future challenges.
- Level Set Segmentation of Cellular Images Based on Topological Dependence24
Topological analysis on the zero level sets is performed to enable effective segmentation of clumped cells in cellular images, able to gain from the advantages of level sets while circumventing its shortcoming.
- Obtaining spatially resolved tumor purity maps using deep multiple instance learning in a pan-cancer study22
A novel deep multiple instance learning model predicting tumor purity from H&E stained digital histopathology slides can be utilized for high throughput sample selection for genomic analysis, which will help reduce pathologists’ workload and decrease inter-observer variability.
- Exploring MRI based radiomics analysis of intratumoral spatial heterogeneity in locally advanced nasopharyngeal carcinoma treated with intensity modulated radiotherapy22
The radiomic features extracted from pre-treatment MRI can potentially reflect the difference between recurrent and non-recurrent regions within a tumor and has a potential role in pre- treatment identification of intra-tumoral radio-resistance for selective dose escalation.
- Enhancing Transformation-Based Defenses Against Adversarial Attacks with a Distribution Classifier22
A separate lightweight distribution classifier is trained to recognize distinct features in the distributions of softmax outputs of transformed images and outperforms majority voting for both clean and adversarial images.
- Reweighting for nonequilibrium Markov processes using sequential importance sampling methods22
This work presents a generic reweighting method for nonequilibrium Markov processes and demonstrates the procedure for the Ising model with the Metropolis algorithm, which can be applied to a variety of systems as well as with different Monte Carlo update schemes.
- 21
- Regional registration of whole slide image stacks containing major histological artifacts21
This paper proposes a new approach to an accurate and robust registration of regions of interest for whole slide images using the idea of multi-scale attention for registration and outperforms the state-of-the-art linear whole tissue registration algorithm.
- Model learning analysis of 3D optoacoustic mesoscopy images for the classification of atopic dermatitis19
A comprehensive analysis using three machine-learning models, random forest (RF), support vector machine (SVM), and convolutional neural network (CNN) for classifying healthy versus AD conditions, and sub-classifying different AD severities using RSOM images and clinical information is conducted.
- Inverse renormalization group based on image super-resolution using deep convolutional networks19
It is demonstrated that the renormalized improved correlation configuration successfully reproduces the original configuration at all the temperatures by the super-resolution scheme.
- 19
- Machine learning improves early prediction of small‐for‐gestational‐age births and reveals nuchal fold thickness as unexpected predictor17
To investigate the performance of the machine learning (ML) model in predicting small‐for‐gestational‐age (SGA) at birth, using second‐trimester data.
- Training machine learning models on patient level data segregation is crucial in practical clinical applications17
It is found that one must be cautious when segregating histological images data (slides) into training, validation and test sets because subtle mishandling of data can introduce data leakage and gives illusively good results on the test set.
- Machine-learning study using improved correlation configuration and application to quantum Monte Carlo simulation16
The Fortuin-Kasteleyn representation-based improved estimator of the correlation configuration is used as an alternative to the ordinary correlation configuration in the machine-learning study of the phase classification of spin models.
- 16
- Reduced representation of protein structure: implications on efficiency and scope of detection of structural similarity15
A reduced representation of protein structure is defined, together with an optimizing function for matching two representations, to provide a pre-filtering stage in a database search and shows that, in a straightforward implementation, the representation performs well in terms of resolution in the space of protein structures, and its ability to make new predictions.
- Automated image based prominent nucleoli detection14
This study developed an accurate prominent nucleoli pattern detector with the potential to be used in the clinical settings and performs twice as good as the use of a single cascade proposed in the seminal paper by Viola and Jones.
- 13
- Systematic benchmarking of 13 AI methods for predicting cyclic peptide membrane permeability12
This study presents the first comprehensive benchmark of 13 machine learning models for predicting the membrane permeability of cyclic peptides, highlighting the superiority of graph-based models, the advantages of regression over classification, and the limitations of scaffold-based data splitting.
- A compact network learning model for distribution regression11
This work designs a compact network representation that encodes and propagates functions in single nodes for the distribution regression task, and achieves higher prediction accuracies while using fewer parameters than traditional neural networks.
- Identification of Cell Nucleus Using a Mumford-Shah Ellipse Detector11
An ellipse detection algorithm based on the Mumford-Shah model that inherits its superior properties and is compared with the randomized Hough transform for detecting nucleus significantly better on data sets.
- 10
- Explaining adversarial vulnerability with a data sparsity hypothesis10
It is found that models trained using this framework, as well as other regularization methods and adversarial training support the hypothesis of data sparsity and that modelstrained with these methods learn to have decision boundaries more similar to the aforementioned ideal decision boundary.
- An AI-assisted tool for efficient prostate cancer diagnosis in low-grade and low-volume cases10
A multi-resolution deep learning pipeline detecting malignant glands in core needle biopsies to help pathologists effectively and accurately diagnose prostate cancer in low-grade and low-volume cases is developed.
- Semi-Supervised subspace learning for Mumford-Shah model based texture segmentation10
A novel semi-supervised optimization algorithm that makes use of information derived from both the intermediate segmentation results and the regions-of-interest selected by the user to determine the optimal subspaces of the target regions.
- Semi-automated quantitative Drosophila wings measurements9
This work has developed a system that automatically detects and measures key points and vein segments on a Drosophila wing and shows that the system performs better than the state of the art.
- Restoration of Uneven Illumination in Light Sheet Microscopy Images9
A modified radiative transfer theory approach is presented to solve the contrast degradation problem of light sheet microscopy (LSM) images and it is confirmed the effectiveness of the approach through simulation as well as real LSM images.
- Segmentation of Neural Stem/Progenitor Cells Nuclei within 3-D Neurospheres9
A novel segmentation approach is presented, called "Evolving Generalized Voronoi Diagram", which uses the identified centers to segment nuclei in neurospheres, and comparison of the computational results to mannually annotated ground truth demonstrates that the proposed approach is an efficient and accurate segmentations approach for 3-D neuro Spheres.
- Analytical Solution to Transport in Brownian Ratchets via the Gambler’s Ruin Model9
An analogy between the classic gambler's ruin problem and the thermally activated dynamics in periodic Brownian ratchets is presented and the conditions for current reversal in the ratchet are obtained.
- Bivariate Nonisotonic Statistical Regression by a Lookup Table Neural System8
The current contribution proposes a neural-network-based data processing method, termed data monotonization, followed by neural isotonic statistical regression, which investigates in particular nonlinear statistical regression of bivariate data that do not exhibit a monotonic dependency.
- A field theoretical restoration method for images degraded by non-uniform light attenuation : an application for light microscopy8
This paper presents a novel physics-based field theoretical approach to solve the contrast degradation problem of light microscopy images and confirms the effectiveness of the technique through simulations as well as through real field experimentations.
- 8
- Exact Conversion of In-Context Learning to Model Weights in Linearized-Attention Transformers7
This paper shows that, for linearized transformer networks, ICL can be made explicit and permanent through the inclusion of bias terms, and mathematically demonstrates the equivalence between a model with ICL demonstration prompts and the same model with the additional bias terms.
- 7
- A Study of Nuclei Classification Methods in Histopathological Images7
This study benchmarks the performance of effective prominent nucleoli detectors in histopathological images along with convolutional and fully connected networks for the task of distinguishing between nuclei with and without prominent nucleolus.
- 7
- Characterizing cardiac adipose tissue in post-acute myocardial infarction patients via CT imaging: a comparative cross-sectional study6
Findings highlight the potential of using CT-derived adipose tissue characteristics to assess inflammation and guide post-AMI management strategies and suggest -70 HU can act as a potential cut-off for inflamed EAT.
- Characterizing Nonculprit Lesions and Perivascular Adipose Tissue of Patients Following Acute Myocardial Infarction Using Coronary Computed Tomography Angiography: A Comparative Study6
Patients post‐AMI displayed heightened noncalcified plaque components, largely due to fibrofatty and necrotic core content, more high‐risk plaques, and increased PVAT mean attenuation on a per‐patient level, highlighting the necessity for refined risk assessment in patients with AMI after treatment.
- 6
- Solving the inverse problem of time independent Fokker–Planck equation with a self supervised neural network method6
This work proposes an FPE-based neural network (FPE-NN) which directly incorporates the FPE terms as neural network weights which allows data-driven scientific discovery of unknown FPE mechanisms.
- Optimal processing for gel electrophoresis images: Applying Monte Carlo Tree Search in GelApp6
The Monte Carlo Tree Search with Upper Confidence Bound (MCTS‐UCB) method is used to efficiently search for optimal image processing pipelines for the band detection task, thereby improving the segmentation algorithm and is a proof‐of‐concept in demonstrating MCTS-UCB as a strategy to optimize general image segmentation.
- Dynamically optimized Wang-Landau sampling with adaptive trial moves and modification factors6
An adaptive variant of the Wang-Landau algorithm is developed that very effectively samples the density of states of continuous models across the entire energy range and performs much better than the traditional Wang- landau sampling.
- Automatic measurement of volume percentage stroma in endometrial images using texture segmentation6
Experimental results show that the proposed Subspace Mumford‐Shah model is well performed on diagnosing premalignant endometrial disease and is very practical for segmenting image set sharing similar properties, which confirm the usefulness of subspace clustering in texture segmentation.
- Automated Protein Distribution Detection in High-Throughput Image-Based siRNA Library Screens6
A computational framework for an automatic detection of changes in images of in vitro cultured keratinocytes when phosphatase genes are silenced using RNAi technology is presented, taking advantage of incorporating prior biological knowledge about phenotypic changes into the algorithm.
- Confusing and Detecting ML Adversarial Attacks with Injected Attractors5
A generic method is given that injects attractors from a watermark decoder into the victim model M, which allows it to leverage on known watermarking schemes for scalability and robustness and provides explainability of the outcomes.
- Accelerated spin dynamics using deep learning corrections5
A Deep Learning method is used to compute the numerical errors of each large time step and use these computed errors to make corrections to achieve higher accuracy in the authors' spin dynamics.
- Confusing and Detecting ML Adversarial Attacks with Injected Attractors through Watermarking5
A generic method is given that injects attractors from a watermark decoder into the victim model M, which allows it to leverage on known watermarking schemes for scalability and robustness and provides explainability of the outcomes.
- Application of Monte Carlo simulation with block-spin transformation based on the Mumford–Shah segmentation model to three-dimensional biomedical images5
The comparison of the output pattern with the clinical experts' annotation suggests that the Mumford-Shah segmentation model is suitable for a multi-phase image segmentationmodel of biomedical images.
- Comparison of multi-label graph cuts method and Monte Carlo simulation with block-spin transformation for the piecewise constant Mumford–Shah segmentation model5
A hybrid method combining the advantages of the Monte Carlo and the graph cuts is proposed that can find the global minimum energy solution efficiently without of initial guess.
- Subspace learning for Mumford–Shah-model-based texture segmentation through texture patches5
A robust and effective algorithm for texture segmentation and feature selection that removes the need to specify training data, which is required by existing methods for the same model and proposes a novel pairwise dissimilarity measure for pixels.
- Monte Carlo methods for optimizing the piecewise constant Mumford–Shah segmentation model5
The Monte Carlo method using several advanced techniques, including block-spin transformation, Eden clustering and simulated annealing, seeks the solution of the celebrated Mumford–Shah image segmentation model.
- Clinicodemographic and Radiological Features of Infective Ring-Enhancing Brain Lesions: A 4-Year Retrospective Study at a Tertiary Referral Center4
Important distinguishing features are revealed between infective REBLs and neoplastic REBLs and between PBAs, TBAs, and NBAs and between PBAs, TBAs, and NBAs.
- Improving transparency and representational generalizability through parallel continual learning4
This is the first effort to train a neural network on multiple tasks and input domains simultaneously in a continual learning scenario and shows that compared to many competing approaches such as continual learning, neural architecture search, and multi-task learning, parallel continual learning is capable of learning more generalizable representations.
- Theoretical and experimental analysis on the generalizability of distribution regression network4
The theoretical properties ofDRN can be used to provide some explanation on the ability of DRN to achieve better generalization performance than conventional neural networks.
- 4
- 4
- Solving the master equation for extremely long time scale calculations4
A new Monte Carlo method for calculating the dynamics of magnetic reversal at arbitrary long time is developed based on microscopic interactions of many constituents and the master equation for magnetic probability distribution function is solved symbolically.
- 3
- Distribution Regression Network.3
The Distribution Regression Network is introduced, which performs regression from input probability distributions to output probability distributions and generalizes the conventional multilayer perceptron (MLP).
- Dynamical traps in Wang-Landau sampling of continuous systems: Mechanism and solution3
A general, simple, and effective solution is proposed where the configurations of multiple parallel Wang-Landau trajectories are interswapped to prevent trapping, explaining why swapping frees the random walker from such traps.
- Region Graph Spectra as Geometric Global Image Features3
A geometric global image feature is developed for pattern retrieval on large bio-image data sets derived by applying spectral graph theory to local feature detectors such as the Scale Invariant Feature Transform, and is effective on patterns with as few as 20 keypoints.
- Deep learning-based quantification of epicardial adipose tissue volume from non-contrast computed tomography images: a multi-centre study2
A deep learning-based system for automated EAT volume quantification using non-contrast computed tomography (NCCT) scans from a large, multi-centre, pan-Asian cohort demonstrated high accuracy and generalizability across ethnically diverse populations, supporting its potential for routine EAT assessment and CAD risk stratification.
- Investigating and unmasking feature-level vulnerabilities of CNNs to adversarial perturbations2
This work introduces a novel framework to study the vulnerability of a CNN model to adversarial perturbations, revealing insights that contribute to a deeper understanding of the phenomenon.
- Super-resolution of spin configurations based on flow-based generative models2
A flow-based generative model that is a deep generative model that is a deep generative model with reversible neural network architecture is presented that can be combined with Monte Carlo simulation to generate large lattice configurations according to the Boltzmann distribution.
- A Weakly Supervised Learning Based Clustering Framework.2
It is mathematically prove that a perfect $ucc$ classifier, in principle, can be used to perfectly cluster individual instances inside the bags, and experimentally shown that the clustering performance of the framework with the classifier is comparable to that of fully supervised learning models.
- 2
- 2
- Accurate single-molecule localization of super-resolution microscopy images using multiscale products2
This work applies the multi-scale product of sub-band images resulting from the wavelet transformation, a technique originally used for astronomical image restoration, for the noise ltering and single-molecule detection in the Super-resolution images.
- A hybrid machine learning framework leveraging biophysicochemical insights for scalable discovery of protein-ligand interactions1
COMRADE (Contrastive Multirepresentation Accelerated Docking Engine), a hybrid virtual screening framework that accelerates docking by triaging compounds using CE-Screen (Contrastive Embedding-Screen), a hybrid virtual screening framework that outperforms state-of-the-art end-to-end models.
- Finding Meaningful Distributions of ML Black-boxes under Forensic Investigation1
This paper proposes leveraging on comprehensive corpus such as ImageNet to select a meaningful distribution that is close to the original training distribution and leads to high performance in follow-up investigations given a poorly documented neural network model.
- 1
- Atopic dermatitis classification models of 3D optoacoustic mesoscopic images1
A comprehensive analysis using three machine-learning models for an AI-aided atopic dermatitis diagnosis and sub-classifying AD severities with 3D Raster Scanning Optoacoustic Mesoscopy images, extracted features from volumetric vascular structures and clinical information.
- Detection and Recovery of Adversarial Attacks with Injected Attractors1
A generic method is given that injects attractors from a watermark decoder into the victim model, which allows it to leverage on known watermarking schemes for scalability and robustness and has competitive performance.
- 1
- Learning Inverse Mappings with Adversarial Criterion.1
A flipped-Adversarial AutoEncoder that simultaneously trains a generative model G that maps an arbitrary latent code distribution to a data distribution and an encoder E that embodies an "inverse mapping" that encodes a data sample into a latent code vector is proposed.
- Predicting time-varying distributions with limited training data1
A new recurrent architecture for DRN is proposed, named recurrent distribution regression network (RDRN), and in experiments involving prediction on sequences of distributions, RDRN and DRN outperform neural network models, with RDRn achieving similar or better accuracies than DRN.
- 1
- An Efficient Network for Predicting Time-Varying Distributions.1
Compared to neural networks and DRN, RDRN achieves the best prediction performance while keeping the network compact, and the combination of compact distribution representation and shared weights architecture across time steps makes the time dependencies in a distribution sequence suitable.
- 1
- Learning Dynamical Shape Prior for Level Set based Cell Tracking1
This paper model the temporal dynamics of shape change using an autoregressive model, which is used for estimating the shape and the location of the current object, and segment the cell using an active contour model starting from the predicted shape.
- 1
- Bit-array alignment effect of perpendicular SOMA media1
The micromagnetic simulation results show that the bit array alignment effect causes large level SNR fluctuation on the same media.
- –
- –
- –
- Abstract P24: Multi-Scale Modeling for Prediction of Spatial Single-Cell and Bulk Transcriptomic Profiles from Whole-Slide Images of Colorectal Cancer–
A graph neural network (GNN) modeling approach is proposed that can model both local interactions between cells within tumor niches, as well as global interactions across tumor regions, and it is shown that the GNN model outperforms existing models for both cell-level and slide-level tasks.
- –
- –
- –
- Oner, Sung, and Lee: Researchers in digital pathology for the future of modern medicine–
A deep learning model for accurate prediction of the proportion of cancer cells within tumor tissue is developed, a necessary step for precision oncology and target therapy in cancer.
- A Deep Learning Network for the Classification of Intracardiac Electrograms in Atrial Tachycardia–
This work elucidates that analysing the EGM signals using a set of explicitly specified rules as proposed by the Decision Trees model is not suitable, and proposes the CNN-LSTM model, which has the ability to learn the complex, intrinsic features within the signals and identify useful features to differentiate the E GM signals.
- Abstract 11812: Effects of Statins on Epicardial Adipose Tissue in Coronary Artery Disease–
Patients on statin treatment showed lower EAT attenuation, with higher EAT HU and percentage of EAT volume within the HU ranges of -60 to -30 and -70 to - 30, which suggests that statin use is associated with reduced EAT inflammation independent of E AT volume.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-10. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it. You will be asked to sign in.