Zhiqiang Shen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhiqiang Shen has a total of 35 co-authors. Here are the top co-authors by number of shared publications:
- Peng Cao
School of Computer Science and Engineering, Northeastern University, Shenyang, China
3 shared publications Jinzhu Yang
2 shared publications- Jinzhu Yang
College of Computer Science and Engineering, Northeastern University, Shenyang, China
2 shared publications Osmar R. Zaiane
Alberta Machine Intelligence Institute, University of Alberta, Edmonton, AB, Canada
2 shared publications- Osmar R. Zaiane
Department of Computing Science, University of Alberta, Edmonton, AB, Canada
2 shared publications Bin Zheng
1 shared publication
Latest Publications
- 20 August 2026
- Published by Elsevier in Artificial Intelligence in Medicine
- Vol. 181
Segmentation-synthesis co-training for semi-supervised domain generalizable medical image segmentation
- 20 August 2026
- Published by Elsevier in Artificial Intelligence in Medicine
- Vol. 181
Segmentation-synthesis co-training for semi-supervised domain generalizable medical image segmentation
- 8 November 2025
- Published by Elsevier in Medical Image Analysis
- Vol. 108, 103857
Adaptive mix for semi-supervised medical image segmentation
- 8 November 2025
- Published by Elsevier in Medical Image Analysis
- Vol. 108, 103857
Adaptive mix for semi-supervised medical image segmentation
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45 (1) , 281-295
Collaborative Learning of Augmentation and Disentanglement for Semi-Supervised Domain Generalized Medical Image Segmentation
This paper explores a challenging yet realistic scenario: semi-supervised domain generalization (SSDG) that includes label scarcity and domain shift problems. We pinpoint that the limitations of previous SSDG methods lie in 1) neglecting the difference between domain shifts existing within a training dataset (intra-domain shift, IDS) and those occurring between training and testing datasets (cross-domain shift, CDS) and 2) overlooking the interplay between label scarcity and domain shifts, resulting in these methods merely stitching together semi-supervised learning (SSL) and domain generalization (DG) techniques. Considering these limitations, we propose a novel perspective to decompose SSDG into the combination of unsupervised domain adaptation (UDA) and DG problems. To this end, we design a causal augmentation and disentanglement framework (CausalAD) for semi-supervised domain generalized medical image segmentation. Concretely, CausalAD involves two collaborative processes: an augmentation process, which utilizes disentangled style factors to perform style augmentation for UDA, and a disentanglement process, which decouples domain-invariant (content) and domain-variant (noise and style) features for DG. Furthermore, we propose a proxy-based self-paced training strategy (ProSPT) to guide the training of CausalAD by gradually selecting unlabeled image pixels with high-quality pseudo labels in a self-paced training manner. Finally, we introduce a hierarchical structural causal model (HSCM) to explain the intuition and concept behind our method. Extensive experiments in the cross-sequence, cross-site, and cross-modality semi-supervised domain generalized medical image segmentation settings show the effectiveness of CausalAD and its superiority over the state-of-the-art. The code is available at https://github.com/Senyh/CausalAD.
- 6 August 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 45 (1) , 281-295
Collaborative Learning of Augmentation and Disentanglement for Semi-Supervised Domain Generalized Medical Image Segmentation
This paper explores a challenging yet realistic scenario: semi-supervised domain generalization (SSDG) that includes label scarcity and domain shift problems. We pinpoint that the limitations of previous SSDG methods lie in 1) neglecting the difference between domain shifts existing within a training dataset (intra-domain shift, IDS) and those occurring between training and testing datasets (cross-domain shift, CDS) and 2) overlooking the interplay between label scarcity and domain shifts, resulting in these methods merely stitching together semi-supervised learning (SSL) and domain generalization (DG) techniques. Considering these limitations, we propose a novel perspective to decompose SSDG into the combination of unsupervised domain adaptation (UDA) and DG problems. To this end, we design a causal augmentation and disentanglement framework (CausalAD) for semi-supervised domain generalized medical image segmentation. Concretely, CausalAD involves two collaborative processes: an augmentation process, which utilizes disentangled style factors to perform style augmentation for UDA, and a disentanglement process, which decouples domain-invariant (content) and domain-variant (noise and style) features for DG. Furthermore, we propose a proxy-based self-paced training strategy (ProSPT) to guide the training of CausalAD by gradually selecting unlabeled image pixels with high-quality pseudo labels in a self-paced training manner. Finally, we introduce a hierarchical structural causal model (HSCM) to explain the intuition and concept behind our method. Extensive experiments in the cross-sequence, cross-site, and cross-modality semi-supervised domain generalized medical image segmentation settings show the effectiveness of CausalAD and its superiority over the state-of-the-art. The code is available at https://github.com/Senyh/CausalAD.
- 13 February 2023
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 155
Learning multi-view and centerline topology connectivity information for pulmonary artery–vein separation
- 13 February 2023
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 155
Learning multi-view and centerline topology connectivity information for pulmonary artery–vein separation
- Open Access
- 10 February 2023
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 21 (1) , 1-12
Association between oxidative stress, mitochondrial function of peripheral blood mononuclear cells and gastrointestinal cancers
Background: The incidence and mortality rate of gastrointestinal cancers are high worldwide. Increasing studies have illustrated that the occurrence, progression, metastasis and prognosis of cancers are intimately linked to the immune system. Mitochondria, as the main source of cellular energy, play an important role in maintaining the physiological function of immune cells. However, the relationship between mitochondrial function of immune cells and tumorigenesis has not yet been systematically investigated. Methods: A total of 150 cases, including 60 healthy donors and 90 primary gastrointestinal cancer patients without anti-tumor treatments (30 with gastric cancer, 30 with liver cancer and 30 with colorectal cancer) were involved in our study. The oxidant/antioxidant and cytokine levels in plasma, the ROS level, mitochondrial function and apoptosis ratio of peripheral blood mononuclear cells (PBMCs) were evaluated. Results: The imbalance between oxidant and antioxidant in plasma was discovered in the primary gastrointestinal cancer patients. The levels of cell reactive oxygen species (ROS) and mitochondrial ROS in PBMCs of primary gastrointestinal cancers were significantly increased compared with that in healthy donors. Meanwhile, the ATP content, the mtDNA copy number and the mitochondrial membrane potential (MMP) in PBMCs of patients with primary gastrointestinal cancers were lower than those in control group. The decreased MMP also occurred in immune cells of gastrointestinal cancers, including T cell, B cell, NK cell and monocyte. Furthermore, the PBMCs apoptosis ratio of primary gastrointestinal cancer patients was significantly higher than that of control group. Importantly, an increase of IL-2 and IL-6 and a decrease of IgG in plasma were found in the patients with primary gastrointestinal cancers. These changes of mitochondrial function in immune cells were consistent among primary gastrointestinal cancers without anti-tumor treatments, such as liver cancer, gastric cancer and colorectal cancer. Conclusion: Our study demonstrated that the imbalance of oxidation/antioxidation in primary gastrointestinal cancer patients without anti-tumor treatments results in excessive ROS. The oxidative stress was associated to the mitochondrial dysfunction, the apoptosis of immune cells and eventually the abnormal immune function in primary gastrointestinal cancers. The application of immune cell mitochondrial dysfunction into clinical evaluation is anticipated.
- 10 February 2023
- Published by Springer Nature in Journal of Translational Medicine
- Vol. 21 (1) , 1-12
Open AccessAssociation between oxidative stress, mitochondrial function of peripheral blood mononuclear cells and gastrointestinal cancers
Background: The incidence and mortality rate of gastrointestinal cancers are high worldwide. Increasing studies have illustrated that the occurrence, progression, metastasis and prognosis of cancers are intimately linked to the immune system. Mitochondria, as the main source of cellular energy, play an important role in maintaining the physiological function of immune cells. However, the relationship between mitochondrial function of immune cells and tumorigenesis has not yet been systematically investigated. Methods: A total of 150 cases, including 60 healthy donors and 90 primary gastrointestinal cancer patients without anti-tumor treatments (30 with gastric cancer, 30 with liver cancer and 30 with colorectal cancer) were involved in our study. The oxidant/antioxidant and cytokine levels in plasma, the ROS level, mitochondrial function and apoptosis ratio of peripheral blood mononuclear cells (PBMCs) were evaluated. Results: The imbalance between oxidant and antioxidant in plasma was discovered in the primary gastrointestinal cancer patients. The levels of cell reactive oxygen species (ROS) and mitochondrial ROS in PBMCs of primary gastrointestinal cancers were significantly increased compared with that in healthy donors. Meanwhile, the ATP content, the mtDNA copy number and the mitochondrial membrane potential (MMP) in PBMCs of patients with primary gastrointestinal cancers were lower than those in control group. The decreased MMP also occurred in immune cells of gastrointestinal cancers, including T cell, B cell, NK cell and monocyte. Furthermore, the PBMCs apoptosis ratio of primary gastrointestinal cancer patients was significantly higher than that of control group. Importantly, an increase of IL-2 and IL-6 and a decrease of IgG in plasma were found in the patients with primary gastrointestinal cancers. These changes of mitochondrial function in immune cells were consistent among primary gastrointestinal cancers without anti-tumor treatments, such as liver cancer, gastric cancer and colorectal cancer. Conclusion: Our study demonstrated that the imbalance of oxidation/antioxidation in primary gastrointestinal cancer patients without anti-tumor treatments results in excessive ROS. The oxidative stress was associated to the mitochondrial dysfunction, the apoptosis of immune cells and eventually the abnormal immune function in primary gastrointestinal cancers. The application of immune cell mitochondrial dysfunction into clinical evaluation is anticipated.
- 25 January 2023
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 154, 106587
WS-LungNet: A two-stage weakly-supervised lung cancer detection and diagnosis network
- 25 January 2023
- Published by Elsevier in Computers in Biology and Medicine
- Vol. 154, 106587
WS-LungNet: A two-stage weakly-supervised lung cancer detection and diagnosis network
- Open Access
Adversarial-Based Knowledge Distillation for Multi-Model Ensemble and Noisy Data Refinement
Generic Image recognition is a fundamental and fairly important visual problem in computer vision. One of the major challenges of this task lies in the fact that single image usually has multiple objects inside while the labels are still one-hot, another one is noisy and sometimes missing labels when annotated by humans. In this paper, we focus on tackling these challenges accompanying with two different image recognition problems: multi-model ensemble and noisy data recognition with a unified framework. As is well-known, usually the best performing deep neural models are ensembles of multiple base-level networks, as it can mitigate the variation or noise containing in the dataset. Unfortunately, the space required to store these many networks, and the time required to execute them at runtime, prohibit their use in applications where test sets are large (e.g., ImageNet). In this paper, we present a method for compressing large, complex trained ensembles into a single network, where the knowledge from a variety of trained deep neural networks (DNNs) is distilled and transferred to a single DNN. In order to distill diverse knowledge from different trained (teacher) models, we propose to use adversarial-based learning strategy where we define a block-wise training loss to guide and optimize the predefined student network to recover the knowledge in teacher models, and to promote the discriminator network to distinguish teacher vs. student features simultaneously. Extensive experiments on CIFAR-10/100, SVHN, ImageNet and iMaterialist Challenge Dataset demonstrate the effectiveness of our MEAL method. On ImageNet, our ResNet-50 based MEAL achieves top-1/5 21.79%/5.99% val error, which outperforms the original model by 2.06%/1.14%. On iMaterialist Challenge Dataset, our MEAL obtains a remarkable improvement of top-3 1.15% (official evaluation metric) on a strong baseline model of ResNet-101.
Open AccessAdversarial-Based Knowledge Distillation for Multi-Model Ensemble and Noisy Data Refinement
Generic Image recognition is a fundamental and fairly important visual problem in computer vision. One of the major challenges of this task lies in the fact that single image usually has multiple objects inside while the labels are still one-hot, another one is noisy and sometimes missing labels when annotated by humans. In this paper, we focus on tackling these challenges accompanying with two different image recognition problems: multi-model ensemble and noisy data recognition with a unified framework. As is well-known, usually the best performing deep neural models are ensembles of multiple base-level networks, as it can mitigate the variation or noise containing in the dataset. Unfortunately, the space required to store these many networks, and the time required to execute them at runtime, prohibit their use in applications where test sets are large (e.g., ImageNet). In this paper, we present a method for compressing large, complex trained ensembles into a single network, where the knowledge from a variety of trained deep neural networks (DNNs) is distilled and transferred to a single DNN. In order to distill diverse knowledge from different trained (teacher) models, we propose to use adversarial-based learning strategy where we define a block-wise training loss to guide and optimize the predefined student network to recover the knowledge in teacher models, and to promote the discriminator network to distinguish teacher vs. student features simultaneously. Extensive experiments on CIFAR-10/100, SVHN, ImageNet and iMaterialist Challenge Dataset demonstrate the effectiveness of our MEAL method. On ImageNet, our ResNet-50 based MEAL achieves top-1/5 21.79%/5.99% val error, which outperforms the original model by 2.06%/1.14%. On iMaterialist Challenge Dataset, our MEAL obtains a remarkable improvement of top-3 1.15% (official evaluation metric) on a strong baseline model of ResNet-101.