Zhen Zhao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Zhen Zhao has a total of 63 co-authors. Here are the top co-authors by number of shared publications:
- Luping Zhou
University of Sydney, Sydney, Australia
6 shared publications - Sifan Long
Jilin University
4 shared publications - Zicheng Wang
University College London, London, United Kingdom and Shanghai AI Laborotory, Shanghai, China
4 shared publications - Jingdong Wang
Baidu, Beijing, China
3 shared publications - Longyue Wang
Alibaba International Digital Commerce, Hangzhou, China
3 shared publications - Shengsheng Wang
College of Computer Science and Technology, Key Laboratory of Symbolic Computation and Knowledge Engineering of Ministry of Education, Jilin University, Changchun, China
3 shared publications
Latest Publications
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6472-6481
Q-Flow: Power-Balance-Aware Graph Learning for Power Flow Calculation with Partially Observed Reactive Power
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6472-6481
Open AccessQ-Flow: Power-Balance-Aware Graph Learning for Power Flow Calculation with Partially Observed Reactive Power
- Open Access
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6482-6491
SenseFlow: A Physics-Informed and Self-Ensembling Iterative Framework for Power Flow Estimation
Power flow estimation plays a vital role in ensuring the stability and reliability of electrical power systems, particularly in the context of growing network complexities and renewable energy integration. However, existing studies often fail to adequately address the unique characteristics of power systems, such as the sparsity of network connections and the critical importance of the unique Slack node, which poses significant challenges in achieving high-accuracy estimations. In this paper, we present SenseFlow, a novel physics-informed and self-ensembling iterative framework that integrates two main designs, the Physics-informed Power Flow Network (PPFNet) and Self-ensembling Iterative Estimation (SeIter), to carefully address the unique properties of the power system and thereby enhance the power flow estimation. In specific, SenseFlow enforces the PPFNet to gradually predict high-precision voltage magnitudes and phase angles through the iterative SeIter process. On the one hand, PPFNet employs the Virtual Node Attention and Slack-Gated Feed-Forward modules to facilitate efficient global-local communication in the face of network sparsity and amplify the influence of the Slack node on angle predictions, respectively. On the other hand, SeIter maintains an exponential moving average of PPFNet's parameters to create an ensemble model that refines power state predictions throughout the iterative fitting process. Experimental results demonstrate that SenseFlow outperforms existing methods, providing a promising solution for high-accuracy power flow estimation across diverse grid configurations. The code is available at https://github.com/ZhenZHAO/SenseFlow.
- 6 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 6482-6491
Open AccessSenseFlow: A Physics-Informed and Self-Ensembling Iterative Framework for Power Flow Estimation
Power flow estimation plays a vital role in ensuring the stability and reliability of electrical power systems, particularly in the context of growing network complexities and renewable energy integration. However, existing studies often fail to adequately address the unique characteristics of power systems, such as the sparsity of network connections and the critical importance of the unique Slack node, which poses significant challenges in achieving high-accuracy estimations. In this paper, we present SenseFlow, a novel physics-informed and self-ensembling iterative framework that integrates two main designs, the Physics-informed Power Flow Network (PPFNet) and Self-ensembling Iterative Estimation (SeIter), to carefully address the unique properties of the power system and thereby enhance the power flow estimation. In specific, SenseFlow enforces the PPFNet to gradually predict high-precision voltage magnitudes and phase angles through the iterative SeIter process. On the one hand, PPFNet employs the Virtual Node Attention and Slack-Gated Feed-Forward modules to facilitate efficient global-local communication in the face of network sparsity and amplify the influence of the Slack node on angle predictions, respectively. On the other hand, SeIter maintains an exponential moving average of PPFNet's parameters to create an ensemble model that refines power state predictions throughout the iterative fitting process. Experimental results demonstrate that SenseFlow outperforms existing methods, providing a promising solution for high-accuracy power flow estimation across diverse grid configurations. The code is available at https://github.com/ZhenZHAO/SenseFlow.
- Open Access
- 10 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 21 (11) , 1-19
SOEDiff: Efficient Distillation for Small Object Editing
In this article, we delve into a new task known as Small Object Editing (SOE), which focuses on text-based image inpainting within a constrained, small-sized area. Despite the remarkable success have been achieved by current image inpainting approaches, their application to the SOE task generally results in failure cases such as Object Missing, Text-Image Mismatch, and Distortion . These failures stem from the limited use of small-sized objects in training datasets and the down-sampling operations employed by U-Net models, which hinders accurate generation. To overcome these challenges, we introduce a novel training-based approach, SOEDiff, aimed at enhancing the capability of baseline models like StableDiffusion in editing small-sized objects while minimizing training costs. Specifically, our method involves two key components: SO-LoRA , which efficiently fine-tunes low-rank matrices, and Cross-scale score distillation , which leverages high-resolution predictions from the pre-trained teacher diffusion model. Our method presents significant improvements on the test dataset collected from MSCOCO and OpenImage, validating the effectiveness of our proposed method in SOE. In particular, when comparing SOEDiff with SD-I model on the OpenImage-small-val dataset, we observe a 0.99 improvement in CLIP-Score and a reduction of 2.87 in FID.
- 10 November 2025
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Multimedia Computing, Communications, and Applications
- Vol. 21 (11) , 1-19
Open AccessSOEDiff: Efficient Distillation for Small Object Editing
In this article, we delve into a new task known as Small Object Editing (SOE), which focuses on text-based image inpainting within a constrained, small-sized area. Despite the remarkable success have been achieved by current image inpainting approaches, their application to the SOE task generally results in failure cases such as Object Missing, Text-Image Mismatch, and Distortion . These failures stem from the limited use of small-sized objects in training datasets and the down-sampling operations employed by U-Net models, which hinders accurate generation. To overcome these challenges, we introduce a novel training-based approach, SOEDiff, aimed at enhancing the capability of baseline models like StableDiffusion in editing small-sized objects while minimizing training costs. Specifically, our method involves two key components: SO-LoRA , which efficiently fine-tunes low-rank matrices, and Cross-scale score distillation , which leverages high-resolution predictions from the pre-trained teacher diffusion model. Our method presents significant improvements on the test dataset collected from MSCOCO and OpenImage, validating the effectiveness of our proposed method in SOE. In particular, when comparing SOEDiff with SD-I model on the OpenImage-small-val dataset, we observe a 0.99 improvement in CLIP-Score and a reduction of 2.87 in FID.
- Open Access
Identification of 4876 Bent-Tail Radio Galaxies in the FIRST Survey Using Deep Learning Combined with Visual Inspection
Bent-tail radio galaxies (BTRGs) are characterized by bent radio lobes. This unique shape is mainly caused by the movement of the galaxy within a cluster, during which the radio jets are deflected by the intra-cluster medium. A combined method, which involves a deep learning-based radio source finder along with visual inspection, has been utilized to search for BTRGs from the Faint Images of the Radio Sky at Twenty-centimeters survey images. Consequently, a catalog of 4876 BTRGs has been constructed, among which 3871 are newly discovered. Based on the classification scheme of the opening angle between the two jets of the galaxy, BTRGs are typically classified as either wide-angle-tail (WAT) sources or narrow-angle-tail (NAT) sources. Our catalog comprises 4424 WATs and 652 NATs. Among these, optical counterparts are identified for 4193 BTRGs. This catalog covers luminosities in the range of $1.91\times10^{20} \leq L_{\rm 1.4\,GHz} \leq 1.45\times10^{28}$ ${\rm W\,Hz^{-1}}$ and redshifts from $z = 0.0023$ to $z = 3.43$. Various physical properties of these BTRGs and their statistics are presented. Particularly, by the nearest neighbor method, we found that 1825 BTRGs in this catalog belong to galaxy clusters reported in literature.Open AccessIdentification of 4876 Bent-Tail Radio Galaxies in the FIRST Survey Using Deep Learning Combined with Visual Inspection
Bent-tail radio galaxies (BTRGs) are characterized by bent radio lobes. This unique shape is mainly caused by the movement of the galaxy within a cluster, during which the radio jets are deflected by the intra-cluster medium. A combined method, which involves a deep learning-based radio source finder along with visual inspection, has been utilized to search for BTRGs from the Faint Images of the Radio Sky at Twenty-centimeters survey images. Consequently, a catalog of 4876 BTRGs has been constructed, among which 3871 are newly discovered. Based on the classification scheme of the opening angle between the two jets of the galaxy, BTRGs are typically classified as either wide-angle-tail (WAT) sources or narrow-angle-tail (NAT) sources. Our catalog comprises 4424 WATs and 652 NATs. Among these, optical counterparts are identified for 4193 BTRGs. This catalog covers luminosities in the range of $1.91\times10^{20} \leq L_{\rm 1.4\,GHz} \leq 1.45\times10^{28}$ ${\rm W\,Hz^{-1}}$ and redshifts from $z = 0.0023$ to $z = 3.43$. Various physical properties of these BTRGs and their statistics are presented. Particularly, by the nearest neighbor method, we found that 1825 BTRGs in this catalog belong to galaxy clusters reported in literature. - Open Access
- 16 January 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 276 (2) , 46
Identification of 4876 Bent-tail Radio Galaxies in the FIRST Survey Using Deep Learning Combined with Visual Inspection
Bent-tail radio galaxies (BTRGs) are characterized by bent radio lobes. This unique shape is mainly caused by the movement of the galaxy within a cluster, during which the radio jets are deflected by the intracluster medium. A combined method, which involves a deep learning-based radio source finder along with visual inspection, has been utilized to search for BTRGs from the Faint Images of the Radio Sky at Twenty cm survey images. Consequently, a catalog of 4876 BTRGs has been constructed, among which 3871 are newly discovered. Based on the classification scheme of the opening angle between the two jets of the galaxy, BTRGs are typically classified as either wide-angle-tail (WAT) sources or narrow-angle-tail (NAT) sources. Our catalog comprises 4424 WATs and 652 NATs. Among these, optical counterparts are identified for 4193 BTRGs. This catalog covers luminosities in the range of 1.91 × 1020 ≤ L 1.4 GHz ≤ 1.45 × 1028 W Hz−1 and redshifts from z = 0.0023 to z = 3.43. Various physical properties of these BTRGs and their statistics are presented. Particularly, by the nearest neighbor method, we found that 1825 BTRGs in this catalog belong to galaxy clusters reported in literature.
- 16 January 2025
- Published by American Astronomical Society in The Astrophysical Journal Supplement Series
- Vol. 276 (2) , 46
Open AccessIdentification of 4876 Bent-tail Radio Galaxies in the FIRST Survey Using Deep Learning Combined with Visual Inspection
Bent-tail radio galaxies (BTRGs) are characterized by bent radio lobes. This unique shape is mainly caused by the movement of the galaxy within a cluster, during which the radio jets are deflected by the intracluster medium. A combined method, which involves a deep learning-based radio source finder along with visual inspection, has been utilized to search for BTRGs from the Faint Images of the Radio Sky at Twenty cm survey images. Consequently, a catalog of 4876 BTRGs has been constructed, among which 3871 are newly discovered. Based on the classification scheme of the opening angle between the two jets of the galaxy, BTRGs are typically classified as either wide-angle-tail (WAT) sources or narrow-angle-tail (NAT) sources. Our catalog comprises 4424 WATs and 652 NATs. Among these, optical counterparts are identified for 4193 BTRGs. This catalog covers luminosities in the range of 1.91 × 1020 ≤ L 1.4 GHz ≤ 1.45 × 1028 W Hz−1 and redshifts from z = 0.0023 to z = 3.43. Various physical properties of these BTRGs and their statistics are presented. Particularly, by the nearest neighbor method, we found that 1825 BTRGs in this catalog belong to galaxy clusters reported in literature.
- Top Cited Papers
- 13 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (4) , 3031-3048
UniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation
Semi-supervised semantic segmentation (SSS) aims at learning rich visual knowledge from cheap unlabeled images to enhance semantic segmentation capability. Among recent works, UniMatch (Yang et al. 2023) improves its precedents tremendously by amplifying the practice of weak-to-strong consistency regularization. Subsequent works typically follow similar pipelines and propose various delicate designs. Despite the achieved progress, strangely, even in this flourishing era of numerous powerful vision models, almost all SSS works are still sticking to 1) using outdated ResNet encoders with small-scale ImageNet-1 K pre-training, and 2) evaluation on simple Pascal and Cityscapes datasets. In this work, we argue that, it is necessary to switch the baseline of SSS from ResNet-based encoders to more capable ViT-based encoders (e.g., DINOv2) that are pre-trained on massive data. A simple update on the encoder (even using 2× fewer parameters) can bring more significant improvement than careful method designs. Built on this competitive baseline, we present our upgraded and simplified UniMatch V2, inheriting the core spirit of weak-to-strong consistency from V1, but requiring less training cost and providing consistently better results. Additionally, witnessing the gradually saturated performance on Pascal and Cityscapes, we appeal that we should focus on more challenging benchmarks with complex taxonomy, such as ADE20K and COCO datasets.
- 13 January 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. 47 (4) , 3031-3048
Top Cited PapersUniMatch V2: Pushing the Limit of Semi-Supervised Semantic Segmentation
Semi-supervised semantic segmentation (SSS) aims at learning rich visual knowledge from cheap unlabeled images to enhance semantic segmentation capability. Among recent works, UniMatch (Yang et al. 2023) improves its precedents tremendously by amplifying the practice of weak-to-strong consistency regularization. Subsequent works typically follow similar pipelines and propose various delicate designs. Despite the achieved progress, strangely, even in this flourishing era of numerous powerful vision models, almost all SSS works are still sticking to 1) using outdated ResNet encoders with small-scale ImageNet-1 K pre-training, and 2) evaluation on simple Pascal and Cityscapes datasets. In this work, we argue that, it is necessary to switch the baseline of SSS from ResNet-based encoders to more capable ViT-based encoders (e.g., DINOv2) that are pre-trained on massive data. A simple update on the encoder (even using 2× fewer parameters) can bring more significant improvement than careful method designs. Built on this competitive baseline, we present our upgraded and simplified UniMatch V2, inheriting the core spirit of weak-to-strong consistency from V1, but requiring less training cost and providing consistently better results. Additionally, witnessing the gradually saturated performance on Pascal and Cityscapes, we appeal that we should focus on more challenging benchmarks with complex taxonomy, such as ADE20K and COCO datasets.
- 31 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (5) , 2016-2027
Imbalanced Medical Image Segmentation With Pixel-Dependent Noisy Labels
Accurate medical image segmentation is often hindered by noisy labels in training data, due to the challenges of annotating medical images. Prior research works addressing noisy labels tend to make class-dependent assumptions, overlooking the pixel-dependent nature of most noisy labels. Furthermore, existing methods typically apply fixed thresholds to filter out noisy labels, risking the removal of minority classes and consequently degrading segmentation performance. To bridge these gaps, our proposed framework, Collaborative Learning with Curriculum Selection (CLCS), addresses pixel-dependent noisy labels with class imbalance. CLCS advances the existing works by i) treating noisy labels as pixel-dependent and addressing them through a collaborative learning framework, and ii) employing a curriculum dynamic thresholding approach adapting to model learning progress to select clean data samples to mitigate the class imbalance issue, and iii) applying a noise balance loss to noisy data samples to improve data utilization instead of discarding them outright. Specifically, our CLCS contains two modules: Curriculum Noisy Label Sample Selection (CNS) and Noise Balance Loss (NBL). In the CNS module, we designed a two-branch network with discrepancy loss for collaborative learning so that different feature representations of the same instance could be extracted from distinct views and used to vote the class probabilities of pixels. Besides, a curriculum dynamic threshold is adopted to select clean-label samples through probability voting. In the NBL module, instead of directly dropping the suspiciously noisy labels, we further adopt a robust loss to leverage such instances to boost the performance. We verify our CLCS on two benchmarks with different types of segmentation noise. Our method can obtain new state-of-the-art performance in different settings, yielding more than 3% Dice and mIoU improvements. Our code is available at https://github.com/Erjian96/CLCS.git.
- 31 December 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Medical Imaging
- Vol. 44 (5) , 2016-2027
Imbalanced Medical Image Segmentation With Pixel-Dependent Noisy Labels
Accurate medical image segmentation is often hindered by noisy labels in training data, due to the challenges of annotating medical images. Prior research works addressing noisy labels tend to make class-dependent assumptions, overlooking the pixel-dependent nature of most noisy labels. Furthermore, existing methods typically apply fixed thresholds to filter out noisy labels, risking the removal of minority classes and consequently degrading segmentation performance. To bridge these gaps, our proposed framework, Collaborative Learning with Curriculum Selection (CLCS), addresses pixel-dependent noisy labels with class imbalance. CLCS advances the existing works by i) treating noisy labels as pixel-dependent and addressing them through a collaborative learning framework, and ii) employing a curriculum dynamic thresholding approach adapting to model learning progress to select clean data samples to mitigate the class imbalance issue, and iii) applying a noise balance loss to noisy data samples to improve data utilization instead of discarding them outright. Specifically, our CLCS contains two modules: Curriculum Noisy Label Sample Selection (CNS) and Noise Balance Loss (NBL). In the CNS module, we designed a two-branch network with discrepancy loss for collaborative learning so that different feature representations of the same instance could be extracted from distinct views and used to vote the class probabilities of pixels. Besides, a curriculum dynamic threshold is adopted to select clean-label samples through probability voting. In the NBL module, instead of directly dropping the suspiciously noisy labels, we further adopt a robust loss to leverage such instances to boost the performance. We verify our CLCS on two benchmarks with different types of segmentation noise. Our method can obtain new state-of-the-art performance in different settings, yielding more than 3% Dice and mIoU improvements. Our code is available at https://github.com/Erjian96/CLCS.git.
- Open Access
- 3 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Anim-Director: A Large Multimodal Model Powered Agent for Controllable Animation Video Generation
- 3 December 2024
- Published by Association for Computing Machinery (ACM)
- p. 1-11
Open AccessAnim-Director: A Large Multimodal Model Powered Agent for Controllable Animation Video Generation
- 30 October 2024
- Published by Springer Nature
- p. 227-243
Alternate Diverse Teaching for Semi-supervised Medical Image Segmentation
- 30 October 2024
- Published by Springer Nature
- p. 227-243
Alternate Diverse Teaching for Semi-supervised Medical Image Segmentation
- Open Access
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3257-3265
Towards Small Object Editing: A Benchmark Dataset and A Training-Free Approach
A plethora of text-guided image editing methods has recently been developed by leveraging the impressive capabilities of large-scale diffusion-based generative models especially Stable Diffusion. Despite the success of diffusion models in producing high-quality images, their application to small object generation has been limited due to difficulties in aligning cross-modal attention maps between text and these objects. Our approach offers a training-free method that significantly mitigates this alignment issue with local and global attention guidance, enhancing the model's ability to accurately render small objects in accordance with textual descriptions. We detail the methodology in our approach, emphasizing its divergence from traditional generation techniques and highlighting its advantages. What's more important is that we also provide SOEBench (Small Object Editing), a standardized benchmark for quantitatively evaluating text-based small object generation collected from MSCOCO[22] and OpenImage[18]. Preliminary results demonstrate the effectiveness of our method, showing marked improvements in the fidelity and accuracy of small object generation compared to existing models. This advancement not only contributes to the field of AI and computer vision but also opens up new possibilities for applications in various industries where precise image generation is critical.We will release our dataset on our project page: https://soebench.github.io/
- 28 October 2024
- Published by Association for Computing Machinery (ACM)
- p. 3257-3265
Open AccessTowards Small Object Editing: A Benchmark Dataset and A Training-Free Approach
A plethora of text-guided image editing methods has recently been developed by leveraging the impressive capabilities of large-scale diffusion-based generative models especially Stable Diffusion. Despite the success of diffusion models in producing high-quality images, their application to small object generation has been limited due to difficulties in aligning cross-modal attention maps between text and these objects. Our approach offers a training-free method that significantly mitigates this alignment issue with local and global attention guidance, enhancing the model's ability to accurately render small objects in accordance with textual descriptions. We detail the methodology in our approach, emphasizing its divergence from traditional generation techniques and highlighting its advantages. What's more important is that we also provide SOEBench (Small Object Editing), a standardized benchmark for quantitatively evaluating text-based small object generation collected from MSCOCO[22] and OpenImage[18]. Preliminary results demonstrate the effectiveness of our method, showing marked improvements in the fidelity and accuracy of small object generation compared to existing models. This advancement not only contributes to the field of AI and computer vision but also opens up new possibilities for applications in various industries where precise image generation is critical.We will release our dataset on our project page: https://soebench.github.io/