Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks134 from public data
- Video Enhancement with Task-Oriented Flow1,802
T task-oriented flow (TOFlow), a motion representation learned in a self-supervised, task-specific manner, is proposed, which outperforms traditional optical flow on standard benchmarks as well as the Vimeo-90K dataset in three video processing tasks.
- MedMNIST v2 - A large-scale lightweight benchmark for 2D and 3D biomedical image classification1,572
Covering primary data modalities in biomedical images, MedMNIST v2 is designed to perform classification on lightweight 2D and 3D images with various dataset scales (from 100 to 100,000) and diverse tasks (binary/multi-class, ordinal regression, and multi-label).
- Learning and Using the Arrow of Time369
A ConvNet suitable for extended temporal footprints and for class activation visualization, and the effect of artificial cues, such as cinematographic conventions, on learning is studied, which achieves state-of-the-art performance on large-scale real-world video datasets.
- 355
- 242
- A connectomic study of a petascale fragment of human cerebral cortex241
In studying these neurons, it is found that deep layer excitatory cell types can be classified into new subsets, based on structural and connectivity differences, and that chandelier interneurons not only innervate excitatories neuron initial segments as previously described, but also each other’s initial segments.
- Dynamic High-Pass Filtering and Multi-Spectral Attention for Image Super-Resolution138
A dynamic high-pass filtering module that locally applies adaptive filter weights for each spatial location and channel group to preserve high-frequency signals and a matrix multi-spectral channel attention module that predicts the attention map of features decomposed in the frequency domain.
- MitoEM Dataset: Large-Scale 3D Mitochondria Instance Segmentation from EM Images132
The MitoEM dataset is introduced, a 3D mitochondria instance segmentation dataset with two (30μm)3 volumes from human and rat cortices respectively, 3, 600× larger than previous benchmarks.
- Seeing the Arrow of Time116
Good video forwards/backwards classification results are demonstrated on a selection of YouTube video clips, and on natively-captured sequences (with no temporally-dependent video compression), and what motions the models have learned that help discriminate forwards from backwards time are examined.
- Motion microscopy for visualizing and quantifying small motions94
A computational tool is presented that quantifies tiny motions in videos and then visualizes them by producing a new video in which the motions are made large enough to see, leading to the discovery of previously unknown phenomena.
- 88
- Learning to Generate Realistic Noisy Images via Pixel-level Noise-aware Adversarial Training81
Qualitative validation shows that noise generated by PNGAN is highly similar to real noise in terms of intensity and distribution, and an efficient architecture Simple Multi-scale Network (SMNet) as the generator is presented.
- 59
- QuantArt: Quantizing Image Style Transfer Towards High Visual Fidelity55
A new style transfer framework for high visual-fidelity stylization called QunntArt, which fusing the quantized and continuous latent representations of the generated artwork toward the centroids of the real artwork distribution with vector quantization.
- 55
- UnMICST: Deep learning with real augmentation for robust segmentation of highly multiplexed images of human tissues54
Two approaches to improving tissue segmentation that are applicable to multiple deep learning frameworks are demonstrated and cumulatively and substantially improve segmentation on a wide range of tissue types and provide a set of improved segmentation models.
- $R^2$-Tuning: Efficient Image-to-Video Transfer Learning for Video Temporal Grounding52
Reversed Recurrent Tuning ($R^2$-Tuning), a parameter- and memory-efficient transfer learning framework for video temporal grounding, achieves state-of-the-art performance across three VTG tasks (i.e., moment retrieval, highlight detection, and video summarization) on six public benchmarks.
- Learning Causality-inspired Representation Consistency for Video Anomaly Detection52
A causality-inspired representation consistency (CRC) framework to implicitly learn the unobservable causal variables of normality directly from available normal videos and detect abnormal events with the learned representation consistency.
- VizAbility: Enhancing Chart Accessibility with LLM-based Conversational Interaction50
VizAbility is a novel system that enriches chart content navigation with conversational interaction, enabling users to use natural language for querying visual data trends, and adapts to the user’s navigation context for improved response accuracy and facilitates verbal command-based chart navigation.
- 50
- Dense 4D nanoscale reconstruction of living brain tissue48
An integrated optical/machine-learning technology, LIONESS (live information-optimized nanoscopy enabling saturated segmentation) leverages optical modifications to stimulated emission depletion microscopy in comprehensively, extracellularly labeled tissue and previous information on sample structure via machine learning to simultaneously achieve isotropic super-resolution, high signal-to-noise ratio and compatibility with living tissue.
- The impact of ChatGPT and LLMs on medical imaging stakeholders: Perspectives and use cases45
This paper introduces an analytic framework for presenting the complex interactions between LLMs and the broader ecosystem of medical imaging stakeholders, including businesses, insurance entities, governments, research institutions, and hospitals (nicknamed BIGR-H).
- 45
- Automated Measurements of Key Morphological Features of Human Embryos for IVF45
This work automates feature extraction of time-lapse microscopy of human embryos with a machine-learning pipeline of five convolutional neural networks (CNNs) that greatly speeds up the measurement of quantitative, biologically relevant features that may aid in embryo selection.
- 44
- NucMM Dataset: 3D Neuronal Nuclei Instance Segmentation at Sub-Cubic Millimeter Scale42
A novel hybrid-representation learning model is proposed that combines the merits of foreground mask, contour map, and signed distance transform to produce high-quality 3D masks and significantly outperforms state-of-the-art nuclei segmentation approaches.
- Pulse-driven robot: Motion via solitary waves41
This study expands the range of possible applications of nonlinear waves and demonstrates that they offer a new platform to make flexible machines to move and reaches a maximum efficiency when the initiated pulses are solitons.
- RibSeg Dataset and Strong Point Cloud Baselines for Rib Segmentation from CT Scans40
A labeled rib segmentation benchmark, named RibSeg, is developed, including 490 CT scans (11,719 individual ribs) from a public dataset, and a point cloud-based baseline method is designed, achieving state-of-the-art segmentation performance with significant efficiency.
- 38
- 37
- Biomedical image analysis competitions: The state of current participation practice36
An international survey was designed that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences, covering participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics.
- Why is the Winner the Best?34
A multicenter study with all 80 benchmarking competitions that were conducted in the scope of IEEE ISBI 2021 and MICCAI 2021 revealed common characteristics of winning solutions: the reflection of the metrics in the method design and the focus on analyzing and handling failure cases.
- CLIPTrans: Transferring Visual Knowledge with Pre-trained Models for Multimodal Machine Translation33
The merits of this framework, which simply adapts the independently pre-trained multimodal M-CLIP and the multilingual mBART, are demonstrated and consequently push forward the state-of-the-art across standard benchmarks by an average of +2.67 BLEU.
- Immersion Fixation and Staining of Multicubic Millimeter Volumes for Electron Microscopy–Based Connectomics of Human Brain Biopsies30
An immersion fixation, heavy metal staining and tissue processing method that consistently provides excellent ultrastructure throughout human and rodent surgical brain samples of volumes 2 × 2 × 2 mm3 and of volumes up to 37 mm3 with one dimension ≤ 2mm is described.
- RibSeg v2: A Large-Scale Benchmark for Rib Labeling and Anatomical Centerline Extraction28
This work extends their prior dataset (RibSeg) on the binary rib segmentation task to a comprehensive benchmark, named RibSeg v2, and develops a pipeline including deep learning-based methods for rib labeling, and a skeletonization-based method for centerline extraction.
- Current Progress and Challenges in Large-Scale 3D Mitochondria Instance Segmentation27
This paper presents eight top-performing approaches from the MitoEM challenge on mitochondria 3D instance segmentation from electron microscopy images, and proposes a new scoring system that better reflects the correctness of the segmentation results.
- AxonEM Dataset: 3D Axon Instance Segmentation of Brain Cortical Regions27
The AxonEM dataset is introduced, which consists of two 30x30x30 um^3 EM image volumes from the human and mouse cortex, respectively, to provide dense 3D axon instance segmentation, enabling large-scale evaluation of axon reconstruction methods.
- Super-BPD: Super Boundary-to-Pixel Direction for Fast Image Segmentation26
This paper proposes a fast image segmentation method based on a novel super boundary-to-pixel direction (super-BPD) and a customized segmentation algorithm with super-B PD, which are novel informative superpixels with robust direction similarity for fast grouping into segmentation regions.
- Deviation magnification25
The framework, termed Deviation Magnification, takes a still image as input, fits parametric models to objects of interest, computes the geometric deviations, and renders an output image in which the departures from ideal geometries are exaggerated.
- Deep Rib Fracture Instance Segmentation and Classification From CT on the RibFrac Challenge23
The analysis revealed that several top rib fracture detection solutions achieved performance comparable or even better than human experts, and lay a foundation for future research and development in AI-assisted rib fracture diagnosis.
- Learning Vector Quantized Shape Code for Amodal Blastomere Instance Segmentation22
The proposed method would enable accurate measurement of blastomeres in In Vitro Fertilization (IVF) clinics, potentially increasing the IVF success rate and generalizability, and outperforms previous state-of-the-art methods.
- 22
- Two Stream Active Query Suggestion for Active Learning in Connectomics22
An end-to-end active learning framework with the query suggestion method for 3D synapse detection and mitochondria segmentation in connectomics, and an unsupervised one optimized on all raw images to capture diverse image features, which can later be improved by fine-tuning on new labels.
- RNAG: a new Gibbs sampler for predicting RNA secondary structure for unaligned sequences21
A new global structural alignment algorithm, RNAG, is presented to predict consensus secondary structures for unaligned sequences using a blocked Gibbs sampling algorithm, which has a theoretical advantage in convergence time.
- A connectomic resource for neural cataloguing and circuit dissection of the larval zebrafish brain19
A correlated light and electron microscopy dataset from a 7-day-old larval zebrafish, integrating confocal imaging of genetically labeled excitatory and inhibitory neurons with nanometer-resolution serial section EM, demonstrates that this resource enables rigorous hypothesis testing as well as exploratory-driven circuit analysis.
- BiaPy: accessible deep learning on bioimages19
BiaPy is an open-source library and application that streamlines the use of common deep learning approaches for bioimage analysis by providing a ready-to-use environment with customizable templates that facilitate sophisticated bioimage analysis workflows.
- Multiplexed expansion revealing for imaging multiprotein nanostructures in healthy and diseased brain19
Multiplexed expansion revealing (multiExR) is reported, which enables high-fidelity antibody visualization of >20 proteins in the same specimen, over serial rounds of staining and imaging, for nanoscale localisation and antibody visualisation of >20 proteins in the same specimen.
- BlastAssist: a deep learning pipeline to measure interpretable features of human embryos18
The BlastAssist pipeline can measure a comprehensive set of interpretable features of human embryos and either outperform or perform comparably to embryologists and human experts in measuring these features.
- 18
- A Data-Driven Regularization Model for Stereo and Flow18
This paper incorporates this data-driven regularization model for stereo or flow through transferring the shape information of the disparity or flow from semantically matched patches in the training database into a standard Markov Random Field model, inferred with a gradient descent algorithm and learned with a discriminative learning approach.
- The Wood Image Analysis and Dataset (WIAD): Open‐access visual analysis tools to advance the ecological data revolution17
The Wood Image Analysis and Dataset (WIAD) an open-source application including a web-interface to integrate visual analysis of wood samples, such as increment cores, thin sections or x-ray films, basic data processing, and arching of the images and derived data to facilitate transparency and reproducibility in studies using visual characteristics of wood.
- 17
- Texture-based Error Analysis for Image Super-Resolution16
This work uses a texture classifier, which enables it to assign patches with semantic labels, to identify the source of SR errors both globally and locally and hopes this work serves as an initial step for debugging blackbox SR networks.
- TriSAM: Tri-Plane SAM for Zero-Shot Cortical Blood Vessel Segmentation in VEM Images14
This paper introduces the first-in-class public benchmark, BvEM, designed specifically for cortical blood vessel segmentation in vEM images, and develops a zero-shot cortical blood vessel segmentation method named TriSAM, which leverages the powerful segmentation model SAM for 3D segmentation.
- Developmental Stage Classification of Embryos Using Two-Stream Neural Network with Linear-Chain Conditional Random Field14
This work proposes a two-stream model for developmental stage classification and develops a linear-chain conditional random field on top of neural network features extracted from the temporal and image streams to make use of both modalities.
- Detecting Synapse Location and Connectivity by Signed Proximity Estimation and Pruning with Deep Nets14
An algorithm to automatically predict the location as well as the direction of both dyadic and polyadic synapses is presented, which outperforms the existing methods for determining synapses in both rodent and fruit fly brain.
- 13
- 13
- Efficient anatomical labeling of pulmonary tree structures via deep point-graph representation-based implicit fields11
This work introduces graph learning on skeletonized structures, incorporating differentiable feature fusion for improved topology and long-distance context capture and employs an implicit function for efficient conversion of sparse representations into dense reconstructions end-to-end.
- Affordance-Aware Object Insertion via Mask-Aware Dual Diffusion11
This paper defines the affordance-aware object insertion task, which aims to seamlessly insert any object into any scene with various position prompts, and proposes the Mask-Aware Dual Diffusion model, which utilizes a dual-stream architecture to simultaneously denoise the RGB image and the insertion mask.
- VizAbility: Multimodal Accessible Data Visualization with Keyboard Navigation and Conversational Interaction11
VizAbility is introduced, a novel multimodal accessible system that combines keyboard navigation with conventional interaction, enabling individuals with visual impairments to actively engage with and explore data visualizations.
- Multiplexed volumetric CLEM enabled by scFvs provides insights into the cytology of cerebellar cortex10
An approach that uses fluorescent single-chain variable fragments (scFvs) to perform multiplexed detergent-free immunolabeling and volumetric-correlated-light-and-electron-microscopy on the same sample suggests potential for molecularly annotated connectomic data.
- 10
- 9
- Mining Visual Evolution in 21 Years of Web Design9
This paper presents a curated dataset of 21 years of web design, scraped from the Internet Archive, and demonstrates how the data can be modeled with deep neural networks to enable novel design applications, such as predicting the apparent year of a web design.
- 3D Domain Adaptive Instance Segmentation via Cyclic Segmentation GANs8
The proposed CySGAN outperforms pre-trained generalist models, feature-level domain adaptation models, and the baselines that conduct image translation and segmentation sequentially and benchmarked on the task of 3D neuronal nuclei segmentation with annotated electron microscopy images and unlabeled expansion microscope data.
- Decoding of the neural representation of the visual RGB color model8
It is shown that RGB color information can be decoded from EEG data and, with the task-irrelevant paradigm, features can bedecoded across fast changes in VEP stimuli, consistent with the theory of both event-related potential (ERP) and P300 mechanisms.
- Domain-Scalable Unpaired Image Translation via Latent Space Anchoring8
This work proposes a new domain-scalable UNIT method, termed as latent space anchoring, which can be efficiently extended to new visual domains and does not need to fine-tune encoders and decoders of existing domains.
- Edge-colored directed subgraph enumeration on the connectome8
A parallel, general-purpose sub graph enumeration strategy to count motifs in the connectome is proposed and a divide-and-conquer community-based subgraph enumeration technique is introduced that allows for enumeration per brain region and allows for differentiation of edges by types to better reflect the underlying biological properties of the graph.
- Genopathomic profiling identifies signatures for immunotherapy response of lung adenocarcinoma via confounder-aware representation learning8
PITER was found to be a useful biomarker to identify patients of lung adenocarcinoma with both favorable progression-free and overall survival in the immunotherapy cohort, and indicated the biological basis in identifying patients with favorable outcome of immunotherapy.
- SCMM: Calibrating cross-modal representations for text-based person search7
Sew Calibration and Masked Modeling is a unified framework that calibrates cross-modal representations for effective multimodal visual-textual pattern matching and strikes an optimal balance between representation expressiveness and computational efficiency by adopting a training-only decoder design.
- Ultrastructural reconstruction of the endodermal nerve net of Hydra vulgaris7
It is speculated that Hydra's endodermal nerve net operates as a non-synaptic circuit, using a neuroendocrine chemical network to implement its functional operations.
- Structure-Preserving Instance Segmentation via Skeleton-Aware Distance Transform7
A skeleton-aware distance transform is proposed that combines the merits of object skeleton in preserving connectivity and DT in modeling geometric arrangement to represent instances with arbitrary structures and achieves state-of-the-art performance.
- 7
- Instance Segmentation of Unlabeled Modalities via Cyclic Segmentation GAN6
This work proposes a novel Cyclic Segmentation Generative Adversarial Network (CySGAN) that conducts image translation and instance segmentation jointly using a unified framework and introduces additional self-supervised and segmentation-based adversarial objectives to improve the model performance by leveraging unlabeled target domain images.
- 6
- Consistent Recurrent Neural Networks For 3d Neuron Segmentation6
A recurrent network for 3D reconstruction of neurons that sequentially generates binary masks for every object in an image with spatio-temporal consistency is presented and achieved state-of-the-art performance on the SNEMI3D challenge.
- Synapse-Aware Skeleton Generation for Neural Circuits6
This work proposes a synapse-aware skeleton generation strategy to transform the reconstructed volumes into an information-rich yet abstract format on which neuroscientists can perform biological analysis and run simulations and demonstrates the results on three large-scale connectomic datasets.
- Scalable and multiplexed recorders of gene regulation dynamics across weeks5
CytoTape reveals that divergent transcriptional trajectories correlate with transcriptional history and signal integration, and that distinct immediate early genes (IEGs) exhibit complex temporal correlations within single cells, and establishes a versatile platform for scalable and multiplexed analysis of cell physiological processes in vitro and in vivo.
- WASPSYN: A Challenge for Domain Adaptive Synapse Detection in Microwasp Brain Connectomes5
To enable developments of domain adaptive synapse detection methods for large-scale connectomics applications, 14 image volumes from a biologically diverse set of Megaphragma viggianii brain regions originating from three different whole-brain datasets are annotated and the WASPSYN challenge at ISBI 2023 is organized.
- Learning Gaze-aware Compositional GAN from Limited Annotations5
This work proposes a Gaze-aware Compositional GAN that learns to generate annotated facial images from a limited labeled dataset and transfers this model to an unlabeled data domain to take advantage of the diversity it provides.
- Multiplexed volumetric CLEM enabled by antibody derivatives provides new insights into the cytology of the mouse cerebellar cortex5
This work developed an approach that uses small fluorescent single-chain variable fragment (scFv) immuno-probes to perform multiplexed detergent-free immuno-labeling and ssEM on the same samples to document a poorly described cell type in the cerebellum, two types of mossy fiber terminals, and the subcellular localization of one type of ion channel.
- 5
- 5
- See the World, Discover Knowledge: A Chinese Factuality Evaluation for Large Vision Language Models4
This paper introduces the first factuality-based visual question-answering benchmark in Chinese, named ChineseSimpleVQA, aimed at assessing the visual factuality of LVLMs across 8 major topics and 56 subtopics, and decoupling the visual factuality into two parts: seeing the world and discovering knowledge.
- Mapping Alzheimer’s Molecular Pathologies in Large-Scale Connectomics Data: A Publicly Accessible Correlative Microscopy Resource4
A volumetric-correlated-light-and-electron microscopy protocol using fluorescent nanobodies to localize Aβ and pTau within a large-scale connectomics dataset from the hippocampus of the 3xTg AD mouse model is presented, uncovered new localizations of Aβ and pTau.
- The XPRESS Challenge: Xray Projectomic Reconstruction -- Extracting Segmentation with Skeletons4
This task provides volumetric XNH images of cortical white matter axons from the mouse brain along with ground truth annotations for axon trajectories along with two types of ground-truth annotations: a small volume of voxel-wise annotations and a larger volume with skeleton-based annotations.
- A Topological Nomenclature for 3D Shape Analysis in Connectomics4
A novel topological nomenclature system to name objects like the appellation for chemical compounds to promote neuroscience analysis based on their skeletal structures by converting the volumetric representation into the topology-preserving reduced graph to untangle the objects.
- 4
- Frenet–Serret Frame-Based Decomposition for Part Segmentation of 3-D Curvilinear Structures3
A novel Frenet-Serret Frame-based Decomposition is introduced, which decomposes 3D curvilinear structures into a globally smooth continuous curve that captures the overall shape, and a cylindrical primitive that encodes local geometric properties.
- Reconstructing sources location of visual color cortex by the task-irrelevant visual stimuli through machine learning decoding3
It is shown that visual color EEG in V1 can reconstruct endocranial neuronal source location, through the machine learning decoding of channel location, and it is explored to trace the brain activity neural source location of EEG inversion problem and assess its reconstructive possibility.
- 3
- 3
- Learning Deformations with Parallel Transport3
This work aims to provide a generative model that explicitly decouples deformation from appearance, which is fundamentally different from the prior work that focuses on deformation-resilient features or metrics.
- NucleoNet and DropNet: Generalist deep learning models for instance segmentation of nuclei and lipid droplets from electron microscopy images2
C crowdsourced the manual step of labeling nuclei and lipid droplets from complex cellular EM images and trained Panoptic DeepLab (PDL) models on these large, heterogenous annotated datasets as well as on publicly available vEM datasets, yielding high-quality models for nuclei and LDs.
- Developing a Deep Learning Approach for Automated Body Composition Prediction in Newborns Using Ultrasound Images2
This study represents the first demonstration of deep learning for automated human body composition prediction from ultrasound images and lays a critical foundation for a novel ultrasound scanning and interpretation protocol to assess malnutrition.
- Developmental Connectomics of the Mouse Cerebellum2
A mechanistic sequence in which early afferent activity and transient cellular structures guide the selective wiring and expansion of parallel fiber input to Purkinje cells, establishing developmental principles that ensure functional specificity in the mature brain is uncovered.
- Enhancing Newborn Health Assessment: Ultrasound-based Body Composition Prediction Using Deep Learning Techniques2
Deep learning provides a promising approach for predicting body composition with ultrasound and a valuable tool for assessing nutritional status in neonatal care.
- WormID-Bench: A Benchmark for Whole-Brain Activity Extraction in C. elegans2
WormID-Bench is introduced, a large-scale, multi-laboratory dataset comprising 118 worms from five distinct research groups, along with standardized evaluation metrics for detection, identification, and tracking, which reveals that existing computational approaches show substantial room for improvement in sensitivity, specificity, and generalization across diverse experimental conditions.
- MitoEM 2.0: A Benchmark for Challenging 3D Mitochondria Instance Segmentation from EM Images2
MitoEM 2.0 facilitates robust model development and fair comparison across methods while supporting reuse in bioimage analysis, algorithm benchmarking, and teaching, and facilitates robust model development and fair comparison across methods while supporting reuse in bioimage analysis, algorithm benchmarking, and teaching.
- Two-stage error detection to improve electron microscopy image mosaicking2
A two-stage error detection method that firstly uses point-based error detection in combination with a hybrid feature framework to expedite the stitching computation while maintaining high accuracy and following is the second detection of unresolved errors with a newly designed metric of EM stitched image quality assessment (EMSIQA).
- An Out-of-Domain Synapse Detection Challenge for Microwasp Brain Connectomes2
Domain adaptation and generalization approaches were proposed to address similar issues for natural images, which were less evaluated on connectomics data due to a lack of out-of-domain benchmarks.
- 2
- Decoding Human Visual Colour EEG Information Using Machine Learning and Visual Evoked Potentials2
The experimental results show that the VEPs can be used for reliable colour real-time decoding in a new kind of task-irrelevant, simple and fast-stimulus BCI-based experimental paradigm that relies on visual evoked potentials (VEPs) during colour observation.
- 2
- Simultaneous brain-wide single-cell recording resolves spatiotemporal memory architecture1
GLOBE resolves the brain-wide spatiotemporal structure of single-cell activity, revealing that Fos transcriptional dynamics associated with fear learning and memory retrieval are distributed across the brain with region-specific temporal heterogeneity, and that the variance of this structure scales down as the number of sampled cells increases.
- 1
- 1
- Learning Causal Domain-Invariant Temporal Dynamics for Few-Shot Action Recognition1
DITeD is proposed, a generative framework with a two-stage training strategy during pre-training that explicitly model invariant dynamics including temporal dynamic generation and transitions, and the variant visual and domain encoders, and validate that the learned temporal dynamic transition and temporal dynamic generation modules possess transferable qualities.
- Learning Socio-Temporal Graphs for Multi-Agent Trajectory Prediction1
This work introduces a Directed Acyclic Graph-based structure, which is term Socio-Temporal Graph (STG), to explicitly capture pair-wise socio-temporal interactions among a group of people across both space and time.
- Multi-Task Curriculum Learning for Partially Labeled Data1
This work proposes a multi-task curriculum learning method that prioritizes less performing tasks during training by setting different supervised learning frequencies for each task, and outperforms standard approaches on one biomedical and two natural image datasets.
- 1
- 1
- –
- –
- NuGraph: Graph-Based Reasoning over 3D Primitives for Nucleus Segmentation Correction–
NuGraph is proposed, a graph-based reasoning framework that operates over atomic 3D primitives obtained by decomposing erroneous masks that encodes primitive geometry via a 3D point-cloud backbone and performs global relational reasoning through graph attention, capturing inter-primitive dependencies across entire clusters rather than isolated pairs.
- A machine learning approach to using ultrasound for body composition and nutritional status assessment in newborns: a pilot study protocol–
This study will determine the feasibility of integrating ultrasound-based body composition assessment into neonatal clinical workflows as a potential future application, and establish a foundation for artificial intelligence (AI)-enabled analysis to enable whole-body composition estimation from ultrasound images.
- –
- –
- S3-TTA: Scale-Style Selection for Test-Time Augmentation in Biomedical Image Segmentation–
A new TTA framework is proposed, S3-TTA, which selects the suitable image scale and style for each test image based on a transformation consistency metric, and constructs an end-to-end augmentation-segmentation joint-training pipeline to ensure a task-oriented augmentation.
- –
- Anatomics MLT, an AI tool for large scale quantification of ultrastructural traits–
This work developed much improved fixation and embedding protocols for plants to drastically reduce processing artifacts and generated an easy-to-use AI tool for quantitative analysis and visualization of large-scale data sets.
- –
- Real-time Image Deblurring to Improve Throughput of Serial-Section Volume Electron Microscopy for Neural Connectomic Studies–
A machine learning-based, real-time image deblurring tool that eliminates the need for retaking out-of-focus images and improves the overall image acquisition performance of smaller datasets where labor costs, storage costs, and machine rental costs are major project considerations.
- –
- –
- –
- –
- –
- –
- –
- Discovering physics and design trends from visual temporal structures–
This dissertation builds computational models to discover and apply the knowledge of the physics for arrow of time, and the design trends for web pages from image sequences, and investigates the evolution of visual cues and layout in web page design through screenshots over time.
- –
- Smoothness- transferred random field–
A new random field model, smoothness-transfer random field (ST-RF) model, for image modeling that with transferred ground truth smoothness, random field can achieve state-of-the-art results in stereo matching and image denoising on standard benchmark dataset.
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 2026-10-15. 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.