Author name cluster

Z. Jane Wang

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

1 author row

NeurIPS Conference 2024 Conference Paper

AdvAD: Exploring Non-Parametric Diffusion for Imperceptible Adversarial Attacks

Jin Li
Ziqiang He
Anwei Luo
Jian-Fang Hu
Z. Jane Wang
Xiangui Kang

Imperceptible adversarial attacks aim to fool DNNs by adding imperceptible perturbation to the input data. Previous methods typically improve the imperceptibility of attacks by integrating common attack paradigms with specifically designed perception-based losses or the capabilities of generative models. In this paper, we propose Adversarial Attacks in Diffusion (AdvAD), a novel modeling framework distinct from existing attack paradigms. AdvAD innovatively conceptualizes attacking as a non-parametric diffusion process by theoretically exploring basic modeling approach rather than using the denoising or generation abilities of regular diffusion models requiring neural networks. At each step, much subtler yet effective adversarial guidance is crafted using only the attacked model without any additional network, which gradually leads the end of diffusion process from the original image to a desired imperceptible adversarial example. Grounded in a solid theoretical foundation of the proposed non-parametric diffusion process, AdvAD achieves high attack efficacy and imperceptibility with intrinsically lower overall perturbation strength. Additionally, an enhanced version AdvAD-X is proposed to evaluate the extreme of our novel framework under an ideal scenario. Extensive experiments demonstrate the effectiveness of the proposed AdvAD and AdvAD-X. Compared with state-of-the-art imperceptible attacks, AdvAD achieves an average of 99. 9% (+17. 3%) ASR with 1. 34 (-0. 97) $l_2$ distance, 49. 74 (+4. 76) PSNR and 0. 9971 (+0. 0043) SSIM against four prevalent DNNs with three different architectures on the ImageNet-compatible dataset. Code is available at https: //github. com/XianguiKang/AdvAD.

PDF Details DOI

AAAI Conference 2024 Conference Paper

PoseGen: Learning to Generate 3D Human Pose Dataset with NeRF

Mohsen Gholami
Rabab Ward
Z. Jane Wang

This paper proposes an end-to-end framework for generating 3D human pose datasets using Neural Radiance Fields (NeRF). Public datasets generally have limited diversity in terms of human poses and camera viewpoints, largely due to the resource-intensive nature of collecting 3D human pose data. As a result, pose estimators trained on public datasets significantly underperform when applied to unseen out-of-distribution samples. Previous works proposed augmenting public datasets by generating 2D-3D pose pairs or rendering a large amount of random data. Such approaches either overlook image rendering or result in suboptimal datasets for pre-trained models. Here we propose PoseGen, which learns to generate a dataset (human 3D poses and images) with a feedback loss from a given pre-trained pose estimator. In contrast to prior art, our generated data is optimized to improve the robustness of the pre-trained model. The objective of PoseGen is to learn a distribution of data that maximizes the prediction error of a given pre-trained model. As the learned data distribution contains OOD samples of the pre-trained model, sampling data from such a distribution for further fine-tuning a pre-trained model improves the generalizability of the model. This is the first work that proposes NeRFs for 3D human data generation. NeRFs are data-driven and do not require 3D scans of humans. Therefore, using NeRF for data generation is a new direction for convenient user-specific data generation. Our extensive experiments show that the proposed PoseGen improves two baseline models (SPIN and HybrIK) on four datasets with an average 6% relative improvement.

PDF Details DOI

JBHI Journal 2021 Journal Article

Bridging the Gap Between 2D and 3D Contexts in CT Volume for Liver and Tumor Segmentation

Lei Song
Haoqian Wang
Z. Jane Wang

Automatic liver and tumor segmentation remain a challenging topic, which subjects to the exploration of 2D and 3D contexts in CT volume. Existing methods are either only focus on the 2D context by treating the CT volume as many independent image slices (but ignore the useful temporal information between adjacent slices), or just explore the 3D context lied in many little voxels (but damage the spatial detail in each slice). These factors lead an inadequate context exploration together for automatic liver and tumor segmentation. In this paper, we propose a novel full-context convolution neural network to bridge the gap between 2D and 3D contexts. The proposed network can utilize the temporal information along the Z axis in CT volume while retaining the spatial detail in each slice. Specifically, a 2D spatial network for intra-slice features extraction and a 3D temporal network for inter-slice features extraction are proposed separately and then are guided by the squeeze-and-excitation layer that allows the flow of 2D context and 3D temporal information. To address the severe class imbalance issue in the CT volume and meanwhile improve the segmentation performance, a loss function consisting of weighted cross-entropy and jaccard distance is proposed. During the network training, the 2D and 3D contexts are learned jointly in an end-to-end way. The proposed network achieves competitive results on the Liver Tumor Segmentation Challenge (LiTS) and the 3D-IRCADB datasets. This method should be a new promising paradigm to explore the contexts for liver and tumor segmentation.

Details DOI

AAAI Conference 2021 Conference Paper

Towards Universal Physical Attacks on Single Object Tracking

Li Ding
Yongwei Wang
Kaiwen Yuan
Minyang Jiang
Ping Wang
Hua Huang
Z. Jane Wang

Recent studies show that small perturbations in video frames could misguide single object trackers. However, such attacks have been mainly designed for digital-domain videos (i. e. , perturbation on full images), which makes them practically infeasible to evaluate the adversarial vulnerability of trackers in real-world scenarios. Here we made the first step towards physically feasible adversarial attacks against visual tracking in real scenes with a universal patch to camouflage single object trackers. Fundamentally different from physical object detection, the essence of single object tracking lies in the feature matching between the search image and templates, and we therefore specially design the maximum textural discrepancy (MTD), a resolution-invariant and target location-independent feature de-matching loss. The MTD distills global textural information of the template and search images at hierarchical feature scales prior to performing feature attacks. Moreover, we evaluate two shape attacks, the regression dilation and shrinking, to generate stronger and more controllable attacks. Further, we employ a set of transformations to simulate diverse visual tracking scenes in the wild. Experimental results show the effectiveness of the physically feasible attacks on SiamMask and SiamRPN++ visual trackers both in digital and physical scenes.

PDF Details

JBHI Journal 2020 Journal Article

An End-to-End Multi-Task Deep Learning Framework for Skin Lesion Analysis

Lei Song
Jianzhe Lin
Z. Jane Wang
Haoqian Wang

Automatic skin lesion analysis of dermoscopy images remains a challenging topic. In this paper, we propose an end-to-end multi-task deep learning framework for automatic skin lesion analysis. The proposed framework can perform skin lesion detection, classification, and segmentation tasks simultaneously. To address the class imbalance issue in the dataset (as often observed in medical image datasets) and meanwhile to improve the segmentation performance, a loss function based on the focal loss and the jaccard distance is proposed. During the framework training, we employ a three-phase joint training strategy to ensure the efficiency of feature learning. The proposed framework outperforms state-of-the-art methods on the benchmarks ISBI 2016 challenge dataset towards melanoma classification and ISIC 2017 challenge dataset towards melanoma segmentation, especially for the segmentation task. The proposed framework should be a promising computer-aided tool for melanoma diagnosis.

Details DOI

JBHI Journal 2019 Journal Article

Dynamic Graph Theoretical Analysis of Functional Connectivity in Parkinson's Disease: The Importance of Fiedler Value

Jiayue Cai
Aiping Liu
Taomian Mi
Saurabh Garg
Wade Trappe
Martin J. McKeown
Z. Jane Wang

Graph theoretical analysis is a powerful tool for quantitatively evaluating brain connectivity networks. Conventionally, brain connectivity is assumed to be temporally stationary, whereas increasing evidence suggests that functional connectivity exhibits temporal variations during dynamic brain activity. Although a number of methods have been developed to estimate time-dependent brain connectivity, there is a paucity of studies examining the utility of brain dynamics for assessing brain disease states. Therefore, this paper aims to assess brain connectivity dynamics in Parkinson's disease (PD) and determine the utility of such dynamic graph measures as potential components to an imaging biomarker. Resting-state functional magnetic resonance imaging data were collected from 29 healthy controls and 69 PD subjects. Time-varying functional connectivity was first estimated using a sliding windowed sparse inverse covariance matrix. Then, a collection of graph measures, including the Fiedler value, were computed and the dynamics of the graph measures were investigated. The results demonstrated that PD subjects had a lower variability in the Fiedler value, modularity, and global efficiency, indicating both abnormal dynamic global integration and local segregation of brain networks in PD. Autoregressive models fitted to the dynamic graph measures suggested that Fiedler value, characteristic path length, global efficiency, and modularity were all less deterministic in PD. With canonical correlation analysis, the altered dynamics of functional connectivity networks, and particularly dynamic Fiedler value, were shown to be related with disease severity and other clinical variables including age. Similarly, Fiedler value was the most important feature for classification. Collectively, our findings demonstrate altered dynamic graph properties, and in particular the Fiedler value, provide an additional dimension upon which to non-invasively and quantitatively assess PD.

Details DOI

YNICL Journal 2018 Journal Article

Decreased subregional specificity of the putamen in Parkinson's Disease revealed by dynamic connectivity-derived parcellation

Aiping Liu
Sue-Jin Lin
Taomian Mi
Xun Chen
Piu Chan
Z. Jane Wang
Martin J. McKeown

Parkinson's Disease (PD) is associated with decreased ability to perform habitual tasks, relying instead on goal-directed behaviour subserved by different cortical/subcortical circuits, including parts of the putamen. We explored the functional subunits in the putamen in PD using novel dynamic connectivity features derived from resting state fMRI recorded from thirty PD subjects and twenty-eight age-matched healthy controls (HC). Dynamic functional segmentation of the putamina was obtained by determining the correlation between each voxel in each putamen along a moving window and applying a joint temporal clustering algorithm to establish cluster membership of each voxel at each window. Contiguous voxels that had consistent cluster membership across all windows were then considered to be part of a homogeneous functional subunit. As PD subjects robustly had two homogenous clusters in the putamina, we also segmented the putamina in HC into two dynamic clusters for a fair comparison. We then estimated the dynamic connectivity using sliding windowed correlation between the mean signal from the identified homogenous subunits and 56 other predefined cortical and subcortical ROIs. Specifically, the mean dynamic connectivity strength and connectivity deviation were then compared to evaluate subregional differences. HC subjects had significant differences in mean dynamic connectivity and connectivity deviation between the two putaminal subunits. The posterior subunit connected strongly to sensorimotor areas, the cerebellum, as well as the middle frontal gyrus. The anterior subunit had strong mean dynamic connectivity to the nucleus accumbens, hippocampus, amygdala, caudate and cingulate. In contrast, PD subjects had fewer differences in mean dynamic connectivity between subunits, indicating a degradation of subregional specificity. Overall UPDRS III and MoCA scores could be predicted using mean dynamic connectivity strength and connectivity deviation. Side of onset of the disease was also jointly related with functional connectivity features. Our results suggest a robust loss of specificity of mean dynamic connectivity and connectivity deviation in putaminal subunits in PD that is sensitive to disease severity. In addition, altered mean dynamic connectivity and connectivity deviation features in PD suggest that looking at connectivity dynamics offers an additional dimension for assessment of neurodegenerative disorders.

Details DOI

JBHI Journal 2017 Journal Article

Automated Detection and Segmentation of Vascular Structures of Skin Lesions Seen in Dermoscopy, With an Application to Basal Cell Carcinoma Classification

Pegah Kharazmi
Mohammed I. AlJasser
Harvey Lui
Z. Jane Wang
Tim K. Lee

Blood vessels are important biomarkers in skin lesions both diagnostically and clinically. Detection and quantification of cutaneous blood vessels provide critical information toward lesion diagnosis and assessment. In this paper, a novel framework for detection and segmentation of cutaneous vasculature from dermoscopy images is presented and the further extracted vascular features are explored for skin cancer classification. Given a dermoscopy image, we segment vascular structures of the lesion by first decomposing the image using independent component analysis into melanin and hemoglobin components. This eliminates the effect of pigmentation on the visibility of blood vessels. Using k-means clustering, the hemoglobin component is then clustered into normal, pigmented, and erythema regions. Shape filters are then applied to the erythema cluster at different scales. A vessel mask is generated as a result of global thresholding. The segmentation sensitivity and specificity of 90% and 86% were achieved on a set of 500 000 manually segmented pixels provided by an expert. To further demonstrate the superiority of the proposed method, based on the segmentation results, we defined and extracted vascular features toward lesion diagnosis in basal cell carcinoma (BCC). Among a dataset of 659 lesions (299 BCC and 360 non-BCC), a set of 12 vascular features are extracted from the final vessel images of the lesions and fed into a random forest classifier. When compared with a few other state-of-art methods, the proposed method achieves the best performance of 96. 5% in terms of area under the curve (AUC) in differentiating BCC from benign lesions using only the extracted vascular features.

Details DOI

JBHI Journal 2017 Journal Article

Illumination Variation-Resistant Video-Based Heart Rate Measurement Using Joint Blind Source Separation and Ensemble Empirical Mode Decomposition

Juan Cheng
Xun Chen
Lingxi Xu
Z. Jane Wang

Recent studies have demonstrated that heart rate (HR) could be estimated using video data [e. g. , exploring human facial regions of interest (ROIs)] under wellcontrolled conditions. However, in practice, the pulse signals may be contaminated by motions and illumination variations. In this paper, tackling the illumination variation challenge, we propose an illumination-robust framework using joint blind source separation (JBSS) and ensemble empirical mode decomposition (EEMD) to effectively evaluate HR from webcam videos. The framework takes the hypotheses that both facial ROI and background ROI have similar illumination variations. The background ROI is then considered as a noise reference sensor to denoise the facial signals by using the JBSS technique to extract the underlying illumination variation sources. Further, the reconstructed illumination-resisted green channel of the facial ROI is detrended and decomposed into a number of intrinsic mode functions using EEMD to estimate the HR. Experimental results demonstrated that the proposed framework could estimate HR more accurately than the state-of-the-art methods. The Bland-Altman plots showed that it led to better agreement with HR ground truth with the mean bias 1. 15 beats/min (bpm), with 95% limits from -15. 43 to 17. 73 bpm, and the correlation coefficient 0. 53. This study provides a promising solution for realistic noncontact and robust HR measurement applications.

Details DOI

JBHI Journal 2014 Journal Article

A Three-Step Multimodal Analysis Framework for Modeling Corticomuscular Activity With Application to Parkinson’s Disease

Xun Chen
Z. Jane Wang
Martin J. McKeown

Corticomuscular coupling analysis based on multiple datasets such as electroencephalography (EEG) and electromyography (EMG) signals provides a useful tool for understanding human motor control systems. A popular conventional method to assess corticomuscular coupling has been the pair-wise magnitude-squared coherence (MSC) between EEG and concomitant EMG recordings. However, there are certain limitations associated with the MSC, including the difficulty in robustly assessing group inference, only dealing with two types of datasets simultaneously and the biologically implausible assumption of pair-wise interactions. To overcome such limitations, in this paper, we propose assessing corticomuscular coupling by combining multiset canonical correlation analysis (M-CCA) and joint independent component analysis (jICA). The proposed method takes advantage of the M-CCA and jICA to ensure that the extracted components are maximally correlated across multiple datasets and meanwhile statistically independent within each dataset. Simulations were performed to illustrate the performance of the proposed method. We also applied the proposed method to concurrent EEG, EMG, and behavior data collected in a Parkinson's disease (PD) study. The results reveal highly correlated temporal patterns among the three types of signals and corresponding spatial activation patterns. In addition to the expected motor areas, the corresponding spatial activation patterns demonstrate enhanced occipital connectivity in the PD subjects, consistent with previous medical findings.

Details DOI

YNIMG Journal 2012 Journal Article

A multiblock PLS model of cortico-cortical and corticomuscular interactions in Parkinson's disease

Joyce Chiang
Z. Jane Wang
Martin J. McKeown

Electroencephalography (EEG) and simultaneously-recorded electromyography (EMG) data are a means to assess integrity of the functional connection between the cortex and the muscle during movement. EEG-EMG coupling is typically assessed with pair-wise squared coherence, resulting in a small, but statistically-significant coherence between a single EEG and a single EMG channel. However, a means to combine results across subjects is not straightforward with this approach because the exact frequency of maximal EEG–EMG coupling may vary between individuals, and it emphasizes the role of an individual locus in the brain in driving the muscle activity, when interactions between head regions may in fact be more influential on ongoing EMG activity. To deal with these issues, we implemented a multiblock Partial Least Squares (mbPLS) procedure, previously proposed in chemical applications, which incorporates a hierarchical structure into the ordinary two-block PLS often used in neuroimaging studies. In the current implementation, each subject's data features are collected in individual data blocks on a sub-level, while simultaneously aggregating the sub-level information to obtain a super-level group “consensus”. We further extended the mbPLS model to include 3-dimensional matrices: time–frequency–EEG channel and a time–frequency–connection utilizing Partial Directed Coherence (PDC). We applied the proposed method to concurrent EEG and EMG data collected from ten normal subjects and nine patients with mild–moderate Parkinson's disease (PD) performing a dynamic motor task—that of sinusoidal squeezing. The results demonstrate that connections between EEG electrodes, rather than activity at individual electrodes, correspond more closely to ongoing EMG activity. In PD subjects, there was enhanced connectivity to and from occipital regions, likely related to the previously-described enhanced use of visual information during motor performance in this group. The proposed mbPLS framework is a promising technique for performing multi-subject, multi-modal data analysis and it allows for robust group inferences even in the face of large inter-subject variability.

Details DOI

YNIMG Journal 2011 Journal Article

Altered directional connectivity in Parkinson's disease during performance of a visually guided task

Giorgia Tropini
Joyce Chiang
Z. Jane Wang
Edna Ty
Martin J. McKeown

Recent animal studies have suggested that cortical areas may play a greater role in the modulation of abnormal oscillatory activity in Parkinson's disease (PD) than previously recognized. We investigated task and medication-dependent, EEG-based directional cortical connectivity in the θ (4–7Hz), α (8–12Hz), β (13–30Hz) and low γ (31–50Hz) frequency bands in 10 PD subjects and 10 age-matched controls. All subjects performed a visually guided task previously shown to modulate abnormal oscillatory activity in PD subjects. We examined the connectivity in the simultaneously-recorded EEG between 5 electrode regions of interest (fronto-central, left and right sensorimotor, central and occipital) using a sparse, multivariate, autoregressive-based partial directed coherence method. For comparison, we utilized traditional Fourier analysis to evaluate task-dependent frequency spectra modulation in these same regions. While the spectral analysis revealed some overall differences between PD and control subjects, it demonstrated relatively modest changes between regions. In contrast, the partial directed coherence-based analysis revealed multifaceted, regionally and directionally-dependent alterations of connectivity in PD subjects during both movement preparation and execution. Connectivity was particularly altered posteriorly, suggesting abnormalities in visual and visuo-motor processing in PD. Moreover, connectivity measures in the α, β and low γ frequency ranges correlated with motor Unified Parkinson's Disease Rating Scores in PD subjects withdrawn from medication. Levodopa administration only partially restored connectivity, and in some cases resulted in further exacerbation of abnormalities. Our results support the notion that PD is associated with significant alterations in connectivity between brain regions, and that these changes can be non-invasively detected in the EEG using partial directed coherence methods. Thus, the role of EEG to monitor PD may need to be further expanded.

Details DOI

JMLR Journal 2009 Journal Article

Controlling the False Discovery Rate of the Association/Causality Structure Learned with the PC Algorithm

Junning Li
Z. Jane Wang

In real world applications, graphical statistical models are not only a tool for operations such as classification or prediction, but usually the network structures of the models themselves are also of great interest (e.g., in modeling brain connectivity). The false discovery rate (FDR), the expected ratio of falsely claimed connections to all those claimed, is often a reasonable error-rate criterion in these applications. However, current learning algorithms for graphical models have not been adequately adapted to the concerns of the FDR. The traditional practice of controlling the type I error rate and the type II error rate under a conventional level does not necessarily keep the FDR low, especially in the case of sparse networks. In this paper, we propose embedding an FDR-control procedure into the PC algorithm to curb the FDR of the skeleton of the learned graph. We prove that the proposed method can control the FDR under a user-specified level at the limit of large sample sizes. In the cases of moderate sample size (about several hundred), empirical experiments show that the method is still able to control the FDR under the user-specified level, and a heuristic modification of the method is able to control the FDR more accurately around the user-specified level. The proposed method is applicable to any models for which statistical tests of conditional independence are available, such as discrete models and Gaussian models. [abs] [ pdf ][ bib ] &copy JMLR 2009. ( edit, beta )

PDF Details

YNIMG Journal 2008 Journal Article

Dynamic Bayesian network modeling of fMRI: A comparison of group-analysis methods

Junning Li
Z. Jane Wang
Samantha J. Palmer
Martin J. McKeown

Bayesian network (BN) modeling has recently been introduced as a tool for determining the dependencies between brain regions from functional-magnetic-resonance-imaging (fMRI) data. However, studies to date have yet to explore the optimum way for meaningfully combining individually determined BN models to make group inferences. We contrasted the results from three broad approaches: the “virtual-typical- subject” (VTS) approach which pools or averages group data as if they are sampled from a single, hypothetical virtual typical subject; the “individual-structure” (IS) approach that learns a separate BN for each subject, and then finds commonality across the individual structures, and the “common-structure” (CS) approach that imposes the same network structure on the BN of every subject, but allows the parameters to differ across subjects. To explore the effects of these three approaches, we applied them to an fMRI study exploring the motor effect of L-dopa medication on ten subjects with Parkinson's disease (PD), as the profound clinical effects of this medication suggest that fMRI activation in PD subjects after medication should start approaching that of age-matched controls. We found that none of these approaches is generally superior over the others, according to Bayesian-information-criterion (BIC) scores, and that they led to considerably different group-level results. The IS approach was more sensitive to the normalization effect of the L-dopa medication on brain connectivity. However, for the more homogeneous control population, the VTS approach was superior. Group-analysis approaches should be selected carefully with consideration of both statistical and biomedical evidence.

Details DOI

YNIMG Journal 2007 Journal Article

Local Linear Discriminant Analysis (LLDA) for group and region of interest (ROI)-based fMRI analysis

Martin J. McKeown
Junning Li
Xuemei Huang
Mechelle M. Lewis
Seungshin Rhee
K.N. Young Truong
Z. Jane Wang

A post-processing method for group discriminant analysis of fMRI is proposed. It assumes that the fMRI data have been pre-processed and analyzed so that each voxel is given a statistic specifying task-related activation(s), and that individually specific regions of interest (ROIs) have been drawn for each subject. The method then utilizes Local Linear Discriminant Analysis (LLDA) to jointly optimize the individually-specific and group linear combinations of ROIs that maximally discriminates between groups (or between tasks, if using the same subjects). LLDA tries to linearly transform each subject's voxel-based activation statistics within ROIs to a common vector space of ROI combinations, enabling the relative similarity of different subjects' activation to be assessed. We applied the method to data recorded from 10 normal subjects during a motor task expected to activate both cortical and subcortical structures. The proposed method detected activation in multiple cortical and subcortical structures that were not present when the data were analyzed by warping the data to a common space. We suggest that the method be applied to group fMRI data when warping to a common space may be ill-advised, such as examining activation in small subcortical structures susceptible to mis-registration, or examining older or neurological patient populations.

Details DOI