Author name cluster

Ying Wu

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

15 papers

2 author rows

YNIMG Journal 2026 Journal Article

PRMix: Primary Region Mix Augmentation and Benchmark Dataset for Precise Whole Mouse Brain Anatomical Delineation

Kunhao Yuan
Hanan Woods
Ülkü Günar
Digin Dominic
Ying Wu
Zhen Qiu
Seth G.N. Grant

The architecture of the mouse brain shares remarkable similarities with the human brain, making it an essential model for studying brain pathologies, synaptic diversity, and regional specialization. A key step in such studies involves registering molecular images to reference brain atlases, a process hindered by the difficulty of accurately delineating brain regions. Toward this, we have curated a collection of high-resolution, dual-fluorescence microscopy images, termed as dual-fluorescence mouse brain microscopy (DMBM) dataset, complemented by expert annotations of 118 subregions in parasagittal sections. This dataset provides unprecedented insights into the molecular and structural complexity of the mouse brain. However, its full potential for detailed whole-brain analysis is compromised by challenges such as boundary ambiguity and sample scarcity in existing automated segmentation methods, prompting the development of the primary region mix (PRMix) augmentation method. PRMix is specifically designed to expand these datasets, enhance the realism of synthetic data and minimize overlap between adjacent regions. Our approach, together with the curated dataset, achieves superior segmentation performance across the mouse brain compared with existing methods, setting a new benchmark in brain imaging research.

Details DOI

EAAI Journal 2026 Journal Article

Virtual evaluation method of bridge load-bearing capacity based on dynamic load test and intelligent algorithm

Pengzhen Lu
Yuchao Liu
Tao Jin
Ying Wu
Xianglong Zheng
Tong Guo

Accurate and timely assessment of bridge bearing capacity is crucial for ensuring structural safety, maintaining traffic flow, and extending service life. Traditional static load testing suffers from limitations like traffic disruption, long duration, and high costs. To address this, this paper proposes an innovative bridge load-bearing capacity assessment framework integrating dynamic load testing with intelligent algorithms. Specifically, key structural parameters are identified through global sensitivity analysis. Dynamic test data are combined with a Bayesian Ridge Regression model to iteratively update the finite element model via back-calculated input parameters. Based on the validation coefficient approach, this method indirectly predicts theoretical static responses by leveraging relationships between static-dynamic characteristics to obtain equivalent static test results, enabling rapid and intelligent evaluation of bridge load performance. Additionally, dynamic assessment and prediction of bearing capacity are achieved using the adjusted FEM and Feature Mode Decomposition (FMD). Through optimization of input variables and training samples, the model demonstrates high accuracy and robust generalization. Case studies validate the method's effectiveness, showing low cost, minimal traffic disruption, and high safety standards, making it particularly suitable for rapid assessment of medium and small span bridges. This approach provides new insights for bridge operation and maintenance, reducing socio-economic costs while enhancing understanding of bridge performance.

Details DOI

IROS Conference 2025 Conference Paper

RMCC: Rigid Multi-joint Coupled Continuum Structure for Bionic Robots

Zida Zhou
Ying Wu
Zujian Chen
Zetong Bi
Hui Cheng

Continuum robots, inspired by biological structures such as spines and tails, have attracted significant attention due to their flexibility and ability to perform complex tasks in confined and dynamic environments. However, traditional flexible continuum robots often encounter challenges such as non-linearity, hysteresis, and limited load-bearing capacity, which can compromise their precision and effectiveness in practical applications. To address these limitations, this paper presents a novel bionic continuum mechanism: Rigid Multi-joint Coupled Continuum Structure(RMCC), which employs a rigid mechanical transmission mode to couple all joints, achieving coordinated movement of multiple joints. Its rigid structural composition and transmission method provide it with high precision and load capacity. The coordinated motion of the joints endows it with the dexterity of a continuum mechanism, while also enabling efficient and precise control with a minimal number of motors. The modular joint design improves the system’s scalability and adaptability, enabling a wide range of configurations to suit diverse robotic applications. The feasibility and effectiveness of the proposed system are validated through a series of bio-inspired experiments, including lizardlike crawling, falling-cat movement, and adaptive grasping like birds. The experimental results confirm that the RMCC exhibits the flexibility and adaptability of animals, demonstrating its potential for diverse bionic robotics applications.

Details

EAAI Journal 2025 Journal Article

Transformer-based partner dance motion generation

Ying Wu
Zizhao Wu
Chengtao Ji

With the rapid development of information technology, particularly the recent breakthroughs in the field of artificial intelligence, the field of dance education has experienced an unprecedented transformation. Today, an increasing number of dance enthusiasts can experience an efficient and personalized immersive dance learning process via advanced artificial intelligence applications. Compared with the individual practice of solo dance, the learning process of partner dancing is more complex. Partner dancing requires not only a coordinated dance partner but also the abilities of both individuals to accurately follow each other's movements in real time as well as adjust and present the correct dance postures. In this study, we developed an innovative real-time virtual reality dance training framework tailored for interactive entertainment and teaching. Our framework used an enhanced transformer model that could generate partner movement sequences based on user movements. Furthermore, we established a somatosensory virtual reality interaction environment by integrating somatosensory devices to capture user movements in real time. These movement sequences were then fed into a network to produce partner movement sequences by generating virtual characters and creating an interactive dance learning system that can offer real-time feedback from a virtual partner. Our experiments validated the robustness and effectiveness of the proposed sequence generation model. In addition, we incorporated Laban movement analysis to explore the semantic representations of partner dances and devised a comprehensive set of evaluation indicators to assess the quality of the generated partner dance movements. The rigorous testing of each component and actual user testing demonstrated the effectiveness of the system. The system offered users rich interactive feedback and a scientifically grounded approach for learning to dance with a partner. Therefore, this study offers significant prospects for both intelligent dance teaching and human-computer interaction in virtual reality among other applications.

Details DOI

EAAI Journal 2024 Journal Article

Fast evaluation method of post-impact performance of bridges based on dynamic load test data using Gaussian process regression

Pengzhen Lu
Yiheng Ma
Ying Wu
Dengguo Li
Tian Jin
Zhenjia Li
Yangrui Chen

Bridges occasionally suffer from the vehicle or ship collision accidents, leading to structural damage and bridge collapse, resulting in severe consequences such as casualties, ship sinking, and vehicle damage. After such accidents, the performance evaluation of bridge structures is significant for bridge maintenance. The bridge's structural performance should be assessed after a collision with a vehicle or ship before regular traffic is resumed. A gray correlation analysis technique was introduced for the swift and efficient assessment of bridge structural performance following impacts. This method aimed to identify the influential parameters associated with bridge structural performance. Utilizing outcomes from dynamic load tests along with the Gaussian process regression model, adjustments were made to the original finite element analysis model. This refinement facilitated precise scrutiny of structural damage and expedited accurate performance evaluations of the bridge. Subsequently, a practical examination was carried out following a ship collision with the Wanjiang Bridge to validate the viability and precision of the proposed approach. A comparison between performance evaluation outcomes derived from the bridge's structural response to ship collision and actual field test results demonstrated the substantial accuracy and computational efficacy of the suggested technique. The proposed method uses a dynamic load test combined with an intelligent algorithm to replace the static load test, effectively solving the expensive, time-consuming, traffic-impeding static load test problem.

Details DOI

ICRA Conference 2024 Conference Paper

Robust and Energy-Efficient Control for Multi-task Aerial Manipulation with Automatic Arm-switching

Ying Wu
Zida Zhou
Mingxin Wei
Hui Cheng

Aerial manipulation has received increasing research interest with wide applications of drones. To perform specific tasks, robotic arms with various mechanical structures will be mounted on the drone. It results in sudden disturbances to the aerial manipulator when switching the robotic arm or interacting with the environment. Hence, it is challenging to design a generic and robust control strategy adapted to various robotic arms when achieving multi-task aerial manipulation. In this paper, we present a learning-based control algorithm that allows online trajectory optimization and tracking to accomplish various aerial interaction tasks without manual adjustment. The proposed energy-saved trajectory planning approach integrates coupled dynamics model with a single rigid body to generate the energy-efficient trajectory for the aerial manipulator. Addressing the challenges of precise control when performing aerial manipulation tasks, this paper presents a controller based on deep neural networks that classifies and learns accurate forces and moments caused by different robotic arms and interactions. Moreover, the forces arising from robotic arm motions are delicately used as part of the drone’s power to save energy. Extensive real-world experiments demonstrate that the proposed method can adapt to various robotic arms and interactions when performing multi-task aerial manipulation.

Details

NeurIPS Conference 2023 Conference Paper

SynMob: Creating High-Fidelity Synthetic GPS Trajectory Dataset for Urban Mobility Analysis

Yuanshao Zhu
Yongchao Ye
Ying Wu
Xiangyu Zhao
James Yu

Urban mobility analysis has been extensively studied in the past decade using a vast amount of GPS trajectory data, which reveals hidden patterns in movement and human activity within urban landscapes. Despite its significant value, the availability of such datasets often faces limitations due to privacy concerns, proprietary barriers, and quality inconsistencies. To address these challenges, this paper presents a synthetic trajectory dataset with high fidelity, offering a general solution to these data accessibility issues. Specifically, the proposed dataset adopts a diffusion model as its synthesizer, with the primary aim of accurately emulating the spatial-temporal behavior of the original trajectory data. These synthesized data can retain the geo-distribution and statistical properties characteristic of real-world datasets. Through rigorous analysis and case studies, we validate the high similarity and utility between the proposed synthetic trajectory dataset and real-world counterparts. Such validation underscores the practicality of synthetic datasets for urban mobility analysis and advocates for its wider acceptance within the research community. Finally, we publicly release the trajectory synthesizer and datasets, aiming to enhance the quality and availability of synthetic trajectory datasets and encourage continued contributions to this rapidly evolving field. The dataset is released for public online availability https: //github. com/Applied-Machine-Learning-Lab/SynMob.

PDF Details

NeurIPS Conference 2023 Conference Paper

TOA: Task-oriented Active VQA

xiaoying xing
Mingfu Liang
Ying Wu

Knowledge-based visual question answering (VQA) requires external knowledge to answer the question about an image. Early methods explicitly retrieve knowledge from external knowledge bases, which often introduce noisy information. Recently large language models like GPT-3 have shown encouraging performance as implicit knowledge source and revealed planning abilities. However, current large language models can not effectively understand image inputs, thus it remains an open problem to extract the image information and input to large language models. Prior works have used image captioning and object descriptions to represent the image. However, they may either drop the essential visual information to answer the question correctly or involve irrelevant objects to the task-of-interest. To address this problem, we propose to let large language models make an initial hypothesis according to their knowledge, then actively collect the visual evidence required to verify the hypothesis. In this way, the model can attend to the essential visual information in a task-oriented manner. We leverage several vision modules from the perspectives of spatial attention (i. e. , Where to look) and attribute attention (i. e. , What to look), which is similar to human cognition. The experiments show that our proposed method outperforms the baselines on open-ended knowledge-based VQA datasets and presents clear reasoning procedure with better interpretability.

PDF Details

YNIMG Journal 2022 Journal Article

Abnormal oxidative metabolism in the cuprizone mouse model of demyelination: An in vivo NIRS-MRI study

Mada Hashem
Qandeel Shafqat
Ying Wu
Jong M. Rho
Jeff F. Dunn

Disruptions in oxidative metabolism may occur in multiple sclerosis and other demyelinating neurological diseases. The impact of demyelination on metabolic rate is also not understood. It is possible that mitochondrial damage may be associated with many such neurological disorders. To study oxidative metabolism with one model of demyelination, we implemented a novel multimodal imaging technique combining Near-Infrared Spectroscopy (NIRS) and MRI to cuprizone mouse model. The cuprizone model is used to study demyelination and may be associated with inhibition of mitochondrial function. Cuprizone mice showed reduced oxygen extraction fraction (-39.1%, p ≤ 0.001), increased tissue oxygenation (6.4%, p ≤ 0.001), and reduced cerebral metabolic rate of oxygen in cortical gray matter (-62.1%, p ≤ 0.001). These changes resolved after the cessation of cuprizone exposure and partial remyelination. A decrease in hemoglobin concentration (-34.4%, p ≤ 0.001), but no change in cerebral blood flow were also observed during demyelination. The oxidized state of the mitochondrial enzyme, Cytochrome C Oxidase (CCO) increased (46.3%, p ≤ 0.001) while the reduced state decreased (-34.4%, p ≤ 0.05) significantly in cuprizone mice. The total amount of CCO did not change significantly during cuprizone exposure. Total CCO did decline after recovery both in control (-23.1%, p ≤ 0.01) and cuprizone (-28.8%, p ≤ 0.001) groups which may relate to age. A reduction in the magnetization transfer ratio, indicating demyelination, was found in the cuprizone group in the cerebral cortex (-3.2%, p ≤ 0.01) and corpus callosum (-5.5%, p ≤ 0.001). In summary, we were able to detect evidence of altered CCO metabolism during cuprizone exposure, consistent with a mitochondrial defect. We observed increased oxygenation and reduced metabolic rate associated with reduced myelination in the gray and white matter. The novel multimodal imaging technique applied here shows promise for noninvasively assessing parameters associated with oxidative metabolism in both mouse models of neurological disease and for translation to study oxidative metabolism in the human brain.

Details DOI

JBHI Journal 2022 Journal Article

Flexible Brain Transitions Between Hierarchical Network Segregation and Integration Associated With Cognitive Performance During a Multisource Interference Task

Rong Wang
Xiaoli Su
Zhao Chang
Pan Lin
Ying Wu

Cognition involves locally segregated and globally integrated processing. This process is hierarchically organized and linked to evidence from hierarchical modules in brain networks. However, researchers have not clearly determined how flexible transitions between these hierarchical processes are associated with cognitive performance. Here, we designed a multisource interference task (MSIT) and introduced the nested-spectral partition (NSP) method to detect hierarchical modules in brain functional networks. By defining hierarchical segregation and integration across multiple levels, we showed that the MSIT requires higher network segregation in the whole brain and most functional systems but generates higher integration in the control system. Meanwhile, brain networks have more flexible transitions between segregated and integrated configurations in the task state. Crucially, higher functional flexibility in the resting state, less flexibility in the task state and more efficient switching of the brain from resting to task states were associated with better task performance. Our hierarchical modular analysis was more effective at detecting alterations in functional organization and the phenotype of cognitive performance than graph-based network measures at a single level.

Details DOI

YNIMG Journal 2020 Journal Article

Using a multimodal near-infrared spectroscopy and MRI to quantify gray matter metabolic rate for oxygen: A hypothermia validation study

Mada Hashem
Qiong Zhang
Ying Wu
Thomas W. Johnson
Jeff F. Dunn

Non-invasive quantitative imaging of cerebral oxygen metabolism (CMRO2) in small animal models is crucial to understand the role of oxidative metabolism in healthy and diseased brains. In this study, we developed a multimodal method combining near-infrared spectroscopy (NIRS) and MRI to non-invasively study oxygen delivery and consumption in the cortex of mouse and rat models. The term CASNIRS is proposed to the technique that measures CMRO2 with ASL and NIRS. To determine the reliability of this method, CMRO2 values were compared with reported values measured with other techniques. Also, the sensitivity of the CASNIRS technique to detect changes in CMRO2 in the cortex of the animals was assessed by applying a reduction in core temperature, which is known to reduce CMRO2. Cerebral blood flow (CBF) and CMRO2 were measured in five mice and five rats at a core temperature of 37 °C followed by another measurement at 33 °C. CMRO2 was 7. 8 ± 1. 8 and 3. 7 ± 0. 9 (ml/100 g/min, mean ± SD) in mice and rats respectively. These values are in good agreement with reported values measured by 15O PET, 17O NMR, and BOLD fMRI. In hypothermia, we detected a significant decrease of 37% and 32% in CMRO2 in the cortex of mice and rats, respectively. Q10 was calculated to be 3. 2 in mice and 2. 7 in rats. In this study we showed that it is possible to assess absolute values of metabolic correlates such as CMRO2, CBF and oxygen extraction fraction (OEF) noninvasively in living brain of mice and rats by combining NIRS with MRI. This will open new possibilities for studying brain metabolism in patients as well as the many mouse/rat models of brain disorders.

Details DOI

YNIMG Journal 2019 Journal Article

Default mode and visual network activity in an attention task: Direct measurement with intracranial EEG

Jiajia Li
Sharif I. Kronemer
Wendy X. Herman
Hunki Kwon
Jun Hwan Ryu
Christopher Micek
Ying Wu
Jason Gerrard

Dynamic attention states are necessary to navigate the ever changing task demands of daily life. Previous investigations commonly utilize a block paradigm to study sustained and transient changes in attention networks. fMRI investigations have shown that sustained attention in visual block design attention tasks corresponds to decreased signal in the default mode and visual processing networks. While task negative networks are anticipated to decrease during active task engagement, it is unexpected that visual networks would also be suppressed during a visual task where event-related fMRI studies have found transient increases to visual stimuli. To resolve these competing results, the current investigations utilized intracranial EEG to directly interrogate visual and default mode network dynamics during a visual continuous performance task. We used the electrophysiological data to model expected fMRI signals and to maximize interpretation of current results with previous investigations. Results show broadband gamma power decreases in the default mode network, corresponding to previous EEG and fMRI findings. Meanwhile, visual processing regions including the primary visual cortex and fusiform gyrus demonstrate both sustained decreases during task engagement and stimuli-driven transient increases in gamma power. Modeled fMRI based on gamma power reproduces signal decreases reported in the fMRI literature, and emphasizes the insensitivity of fMRI to transient, regularly spaced signal changes embedded within sustained network dynamics. The signal processing functions of the dynamic visual and default mode network changes explored in this study are unknown but may be elucidated through further investigation.

Details DOI

AAAI Conference 2016 Conference Paper

Learning FRAME Models Using CNN Filters

Yang Lu
Song-Chun Zhu
Ying Wu

The convolutional neural network (ConvNet or CNN) has proven to be very successful in many tasks such as those in computer vision. In this conceptual paper, we study the generative perspective of the discriminative CNN. In particular, we propose to learn the generative FRAME (Filters, Random ﬁeld, And Maximum Entropy) model using the highly expressive ﬁlters pre-learned by the CNN at the convolutional layers. We show that the learning algorithm can generate realistic and rich object and texture patterns in natural scenes. We explain that each learned model corresponds to a new CNN unit at a layer above the layer of ﬁlters employed by the model. We further show that it is possible to learn a new layer of CNN units using a generative CNN model, which is a product of experts model, and the learning algorithm admits an EM interpretation with binary latent variables.

PDF Details

YNIMG Journal 2006 Journal Article

Automated segmentation of multiple sclerosis lesion subtypes with multichannel MRI

Ying Wu
Simon K. Warfield
I. Leng Tan
William M. Wells
Dominik S. Meier
Ronald A. van Schijndel
Frederik Barkhof
Charles R.G. Guttmann

Purpose. To automatically segment multiple sclerosis (MS) lesions into three subtypes (i. e. , enhancing lesions, T1 “black holes”, T2 hyperintense lesions). Materials and methods. Proton density-, T2- and contrast-enhanced T1-weighted brain images of 12 MR scans were pre-processed through intracranial cavity (IC) extraction, inhomogeneity correction and intensity normalization. Intensity-based statistical k-nearest neighbor (k-NN) classification was combined with template-driven segmentation and partial volume artifact correction (TDS+) for segmentation of MS lesions subtypes and brain tissue compartments. Operator-supervised tissue sampling and parameter calibration were performed on 2 randomly selected scans and were applied automatically to the remaining 10 scans. Results from this three-channel TDS+ (3ch-TDS+) were compared to those from a previously validated two-channel TDS+ (2ch-TDS+) method. The results of both the 3ch-TDS+ and 2ch-TDS+ were also compared to manual segmentation performed by experts. Results. Intra-class correlation coefficients (ICC) of 3ch-TDS+ for all three subtypes of lesions were higher (ICC between 0. 95 and 0. 96) than that of 2ch-TDS+ for T2 lesions (ICC = 0. 82). The 3ch-TDS+ also identified the three lesion subtypes with high specificity (98. 7–99. 9%) and accuracy (98. 5–99. 9%). Sensitivity of 3ch-TDS+ for T2 lesions was 16% higher than with 2ch-TDS+. Enhancing lesions were segmented with the best sensitivity (81. 9%). “Black holes” were segmented with the least sensitivity (62. 3%). Conclusion. 3ch-TDS+ is a promising method for automated segmentation of MS lesion subtypes.

Details DOI

EAAI Journal 2002 Journal Article

Towards self-exploring discriminating features for visual learning

Ying Wu
Thomas S. Huang

Many visual learning tasks are usually confronted by some common difficulties. One of them is the lack of supervised information, due to the fact that labeling could be tedious, expensive or even impossible. Another difficulty is the high dimensionality of the visual data. Fortunately, these difficulties could be alleviated by using a hybrid of labeled and unlabeled training data for learning. Since the unlabeled data characterize the joint probability across different features, they could be used to boost weak classifiers by exploring discriminating features in a self-supervised fashion. This paper proposes a novel method, the Discriminant-EM (D-EM) algorithm, which attacks these difficulties by integrating discriminant analysis with the EM framework in this hybrid formulation. Both linear and nonlinear methods are investigated in this paper. Based on kernel multiple discriminant analysis, the nonlinear D-EM provides better ability to simplify the probabilistic structures of data distributions in a discrimination space. We also propose a novel data-sampling scheme for efficient learning of kernel discriminants. Our experimental results show that D-EM outperforms a variety of supervised and semi-supervised learning algorithms for many visual learning tasks, such as content-based image retrieval, invariant object recognition, and nonstationary color tracking. The proposed approach could be easily applied for many other learning tasks.

Details DOI