Author name cluster

Qiang Dong

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.

8 papers

1 author row

YNICL Journal 2025 Journal Article

A deep learning approach versus expert clinician panel in the classification of posterior circulation infarction

Leon S. Edwards
Milanka Visser
Cecilia Cappelen-Smith
Dennis Cordato
Andrew Bivard
Leonid Churilov
Christopher Blair
James Thomas

BACKGROUND: Posterior circulation infarction (POCI) is common. Imaging techniques such as non-contrast-CT (NCCT) and diffusion-weighted-magnetic-resonance-imaging commonly fail to detect hyperacute POCI. Studies suggest expert inspection of Computed Tomography Perfusion (CTP) improves diagnosis of POCI. In many settings, there is limited access to specialist expertise. Deep-learning has been successfully applied to automate imaging interpretation. This study aimed to develop and validate a deep-learning approach for the classification of POCI using CTP. METHODS: Data were analysed from 3541-patients from the International-stroke-perfusion-registry (INSPIRE). All patients with baseline multimodal-CT and follow-up imaging performed at 24-48 h were identified. A cohort of 541-patients was constructed on a 1:3 POCI-to -reference-ratio for model analysis. A 3D-Dense-Convolutional-Network (DenseNet) was trained to classify patients into POCI or non-POCI using CTP-deconvolved-maps. Six-stroke-experts also independently classified patients based upon stepwise access to multimodal CT (mCT) data. DenseNet results were compared against expert clinician results. Model and clinician performance was evaluated using area-under-the-receiver-operating-curve, sensitivity, specificity, accuracy and precision. Clinician agreement was measured with the Fleiss-Kappa-statistic. RESULTS: Best mean clinician diagnostic accuracy, sensitivity and agreement was demonstrated after review of all mCT data (AUC: 0.81, Sensitivity: 0.65, Fleiss-Kappa-statistic: 0.73). There was a spectrum of individual clinician results with an AUC-range of 0.73-0.86. Best DenseNet performance was recorded with an input combination of NCCT and delay-time maps. The DenseNet model was superior to the best mean clinician performance (AUC: 0.87) and was due to enhanced sensitivity (DenseNET: 0.77, Clinician: 0.65). The degree to which the DenseNet model outperformed each clinician ranged and was clinician specific (AUC improvement 0.01-0.14). CONCLUSION: Comprehensive review of CTP improves diagnostic performance and agreement amongst clinicians. A DenseNet model was superior to best mean clinician performance. The degree of improvement varied by specific clinician. Development of a clinician-DenseNet approach may improve inter-clinician agreement and diagnostic accuracy. This approach may alleviate limited specialist services in resource constrained settings.

Details DOI

AAAI Conference 2025 Conference Paper

Adaptive Dataset Quantization

Muquan Li
Dongyang Zhang
Qiang Dong
Xiurui Xie
Ke Qin

Contemporary deep learning, characterized by the training of cumbersome neural networks on massive datasets, confronts substantial computational hurdles. To alleviate heavy data storage burdens on limited hardware resources, numerous dataset compression methods such as dataset distillation (DD) and coreset selection have emerged to obtain a compact but informative dataset through synthesis or selection for efficient training. However, DD involves an expensive optimization procedure and exhibits limited generalization across unseen architectures, while coreset selection is limited by its low data keep ratio and reliance on heuristics, hindering its practicality and feasibility. To address these limitations, we introduce a newly versatile framework for dataset compression, namely Adaptive Dataset Quantization (ADQ). Specifically, we first identify the sub-optimal performance of naive Dataset Quantization (DQ), which relies on uniform sampling and overlooks the varying importance of each generated bin. Subsequently, we propose a novel adaptive sampling strategy through the evaluation of generated bins' representativeness score, diversity score and importance score, where the former two scores are quantified by the texture level and contrastive learning-based techniques, respectively. Extensive experiments demonstrate that our method not only exhibits superior generalization capability across different architectures, but also attains state-of-the-art results.

PDF Details DOI

TCS Journal 2024 Journal Article

The diameter of rectangular twisted torus

Qiang Dong
Juan Zhao

The rectangular twisted torus is an attractive alternative to the classic torus as interconnection topology of high-performance computers. In this article, we establish the analytical expression of the diameter of a rectangular twisted torus with any aspect ratio and any twist slope. The result validates the optimality of a rectangular twisted torus with width-height-slope ratio 2: 1: 1.

Details DOI

YNICL Journal 2020 Journal Article

Genome-wide association study of white matter hyperintensity volume in elderly persons without dementia

Yu Guo
Xue-Ning Shen
Xiao-He Hou
Ya-Nan Ou
Yu-Yuan Huang
Qiang Dong
Lan Tan
Jin-Tai Yu

BACKGROUND: White matter hyperintensity has been correlated with cognitive disorders and its genetic predictors remain unclear. Here we conducted a genome-wide association study to identify novel genetic determinants that were correlated with white matter hyperintensity volume (WMHV) among non-demented elders. METHODS: Three hundred and fifty non-Hispanic Caucasian subjects aged 55-80 years were included from the Alzheimer's Disease Neuroimaging Initiative cohort. Associations of WMHV with genetic polymorphisms were explored using multiple linear regression under an additive genetic model. Further studies were conducted to explore the influence of genetic variants on cognition-related phenotypes. RESULTS: ) were identified as suggestive loci linked to WMHV levels. The minor allele of rs7220676 (C) showed association with lower log (WMHV) in a dose-dependent manner. Besides, rs7220676 was correlated with rates of cognitive decline assessed by Mini-mental State Examination and memory scores. CONCLUSIONS: A novel locus near HS3ST3A1 and MIR548H3 genes was associated with WMHV levels and it may be involved in neurodegenerative diseases.

Details DOI

YNICL Journal 2019 Journal Article

Deep/mixed cerebral microbleeds are associated with cognitive dysfunction through thalamocortical connectivity disruption: The Taizhou Imaging Study

Yingzhe Wang
Yanfeng Jiang
Chen Suo
Ziyu Yuan
Kelin Xu
Qi Yang
Weijun Tang
Kexun Zhang

BACKGROUND: Cerebral microbleeds (CMBs) are considered to be risk factors for cognitive dysfunction. The specific pathology and clinical manifestations of CMBs are different based on their locations. We investigated the association between CMBs at different locations and cognitive dysfunction and explored the potential underlying pathways in a rural Han Chinese population. METHODS: We used baseline data from 562 community-dwelling adults (55-65 years old) in the Taizhou Imaging Study between 2013 and 2015. All individuals underwent multimodal brain magnetic resonance imaging (MRI) and 444 subjects completed neuropsychological tests: the Mini-Mental Status Examination and the Montreal Cognitive Assessment. Multinomial logistic regression was used to estimate the association between CMBs and cognitive dysfunction. The volume of brain regions and white matter microstructure were analyzed using Freesurfer and tract-based spatial statistics, respectively. RESULTS: CMBs were detected in 104 individuals (18.5%) in our study. Multinomial logistic regression found deep/mixed CMBs were associated with global cognitive dysfunction (OR 3.52; 95% CI 1.21 to 10.26), whereas lobar CMBs (OR 1.76; 95% CI 0.56 to 5.53) were not. Quantification of multimodal brain MRI showed that deep/mixed CMBs were accompanied by decreased thalamic volume and loss of fractional anisotropy of bilateral anterior thalamic radiations. CONCLUSION: Deep/mixed CMBs were associated with cognitive dysfunction in this Chinese cross-sectional study. Disruption of thalamocortical connectivity might be a potential pathway underlying this relationship.

Details DOI

TCS Journal 2019 Journal Article

How many triangles and quadrilaterals are there in an n-dimensional augmented cube?

Qiang Dong
Xi Wang

The augmented cube is an important variant of hypercube as interconnection topology of parallel computing. In this paper, we examine the numbers of short cycles in augmented cubes, and prove that for n ≥ 3, there are 2 n × ( n − 1 ) triangles and 2 n − 2 × ( 2 n 2 + 5 n − 11 ) quadrilaterals in an n-dimensional augmented cube. This result shows that augmented cubes are promising interconnection networks with superior connectivity and fault-tolerant capability.

Details DOI

TCS Journal 2013 Journal Article

Hamiltonian connectivity of restricted hypercube-like networks under the conditional fault model

Qiang Dong
Junlin Zhou
Yan Fu
Hui Gao

Restricted hypercube-like networks (RHLNs) are an important class of interconnection networks for parallel computing systems, which include most popular variants of the hypercubes, such as crossed cubes, Möbius cubes, twisted cubes and locally twisted cubes. This paper deals with the fault-tolerant hamiltonian connectivity of RHLNs under the conditional fault model. Let G be an n -dimensional RHLN and F ⊆ V ( G ) ⋃ E ( G ), where n ≥ 7 and ∣ F ∣ ≤ 2 n − 10. We prove that for any two nodes u, v ∈ V ( G − F ) satisfying a simple necessary condition on neighbors of u and v, there exists a hamiltonian or near-hamiltonian path between u and v in G − F. The result extends further the fault-tolerant graph embedding capability of RHLNs.

Details DOI

TCS Journal 2011 Journal Article

Hamiltonian properties of twisted hypercube-like networks with more faulty elements

Xiaofan Yang
Qiang Dong
Erjie Yang
Jianqiu Cao

Twisted hypercube-like networks (THLNs) are a large class of network topologies, which subsume some well-known hypercube variants. This paper is concerned with the longest cycle in an n -dimensional ( n -D) THLN with up to 2 n − 9 faulty elements. Let G be an n -D THLN, n ≥ 7. Let F be a subset of V ( G ) ⋃ E ( G ), | F | ≤ 2 n − 9. We prove that G − F contains a Hamiltonian cycle if δ ( G − F ) ≥ 2, and G − F contains a near Hamiltonian cycle if δ ( G − F ) ≤ 1. Our work extends some previously known results.

Details DOI