Fanhua Shang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Fanhua Shang has a total of 172 co-authors. Here are the top co-authors by number of shared publications:
- Hongying Liu
Medical School, Tianjin University, Tianjin, China
34 shared publications - Yuanyuan Liu
School of Artificial Intelligence, Xidian University, Xi'an, China
27 shared publications - Licheng Jiao
Xidian University, Xian, China
12 shared publications - Wei Feng
School of Computer Science and Technology, Tianjin University, Tianjin, China
12 shared publications - Liang Wan
Tianjin University, Tianjin, China
8 shared publications - Fan Lyu
Institute of Automation, Chinese Academy of Sciences
7 shared publications
Latest Publications
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9638-9647
MLLM-Facilitated Peer-Review Self-Correction for Continual Domain Shift Learning
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 9638-9647
Open AccessMLLM-Facilitated Peer-Review Self-Correction for Continual Domain Shift Learning
- Open Access
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 12526-12534
DGSRF: Dual-Text Guided Super-Resolution Fusion via Interactive Gating and Hybrid Instruction Modulation
- 9 November 2026
- Published by Association for Computing Machinery (ACM)
- p. 12526-12534
Open AccessDGSRF: Dual-Text Guided Super-Resolution Fusion via Interactive Gating and Hybrid Instruction Modulation
- 25 September 2026
- Published by Springer Nature
- p. 57-66
Breaking Teacher’s Mean Regression Bias: Boundary-Aware Contrastive Distillation for Endoscopic Super-Resolution
- 25 September 2026
- Published by Springer Nature
- p. 57-66
Breaking Teacher’s Mean Regression Bias: Boundary-Aware Contrastive Distillation for Endoscopic Super-Resolution
- Open Access
- 7 June 2026
- Published by Elsevier in SSRN Electronic Journal
A Review of Deep Learning-based Super-Resolution Methods for Medical Imagery Sequences
Medical image super-resolution (SR) aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs and plays a critical role in enhancing diagnostic accuracy and treatment planning. While previous surveys have extensively covered single-image SR techniques, comparatively little attention has been paid to the emerging domain of multi-slice or sequence-based medical image SR, which exploits spatial and temporal correlations across adjacent slices or frames. This paper presents the first comprehensive survey dedicated to super-resolution of medical image sequences. We introduce a novel stereoscopic taxonomy structured as a dual-dimensional classification framework, categorizing existing methods according to two key aspects: (1) input sequence types (static image sequences or dynamic video frames) and (2) imaging spatial dimensions (in-plane, through-plane, and volumetric). This taxonomy offers a unified perspective on SR techniques applied across diverse medical imaging modalities, including magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, and endoscopy. Furthermore, we summarize widely used degradation models, image quality assessment metrics, and benchmark datasets. Finally, we systematically evaluate the strengths and limitations of state-of-the-art methods and highlight emerging trends and open challenges in the field.
- 7 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessA Review of Deep Learning-based Super-Resolution Methods for Medical Imagery Sequences
Medical image super-resolution (SR) aims to reconstruct high-resolution (HR) images from low-resolution (LR) inputs and plays a critical role in enhancing diagnostic accuracy and treatment planning. While previous surveys have extensively covered single-image SR techniques, comparatively little attention has been paid to the emerging domain of multi-slice or sequence-based medical image SR, which exploits spatial and temporal correlations across adjacent slices or frames. This paper presents the first comprehensive survey dedicated to super-resolution of medical image sequences. We introduce a novel stereoscopic taxonomy structured as a dual-dimensional classification framework, categorizing existing methods according to two key aspects: (1) input sequence types (static image sequences or dynamic video frames) and (2) imaging spatial dimensions (in-plane, through-plane, and volumetric). This taxonomy offers a unified perspective on SR techniques applied across diverse medical imaging modalities, including magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, and endoscopy. Furthermore, we summarize widely used degradation models, image quality assessment metrics, and benchmark datasets. Finally, we systematically evaluate the strengths and limitations of state-of-the-art methods and highlight emerging trends and open challenges in the field.
- 4 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. PP (01628828) , 1-18
Elastic Multi-Gradient Descent for Parallel Continual Learning
Learning a single shared model across multiple tasks is a long-standing goal in machine learning, traditionally addressed through Multi-Task Learning (MTL) or Serial Continual Learning (SCL). However, MTL assumes fixed task availability, limiting adaptability to new tasks, while SCL suffers from delayed response due to its strictly sequential learning protocol. This paper studies the novel paradigm of Parallel Continual Learning (PCL) in dynamic multi-task scenarios, where a diverse set of tasks is encountered at different time points. PCL inherits the challenges of both MTL and SCL, namely, task conflict and catastrophic forgetting, compounded by the dynamic and asynchronous arrival of tasks. To address these challenges, we propose Elastic Multi-Gradient Descent (EMGD), which formulates PCL as a dynamic multi-objective optimization problem. EMGD introduces task-specific elastic factors to guide gradient updates toward Pareto-optimal directions, ensuring balanced learning across tasks. Additionally, we develop a gradient-guided memory editing mechanism that aligns rehearsal data with the optimized descent direction, mitigating memory-induced interference. Theoretical analysis shows that accumulation points of the EMGD update rule are Pareto critical under the proposed formulation, and extensive experiments on image classification benchmarks demonstrate that EMGD significantly outperforms existing methods, including state-of-the-art PCL, MTL, and SCL approaches.
- 4 May 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Pattern Analysis and Machine Intelligence
- Vol. PP (01628828) , 1-18
Elastic Multi-Gradient Descent for Parallel Continual Learning
Learning a single shared model across multiple tasks is a long-standing goal in machine learning, traditionally addressed through Multi-Task Learning (MTL) or Serial Continual Learning (SCL). However, MTL assumes fixed task availability, limiting adaptability to new tasks, while SCL suffers from delayed response due to its strictly sequential learning protocol. This paper studies the novel paradigm of Parallel Continual Learning (PCL) in dynamic multi-task scenarios, where a diverse set of tasks is encountered at different time points. PCL inherits the challenges of both MTL and SCL, namely, task conflict and catastrophic forgetting, compounded by the dynamic and asynchronous arrival of tasks. To address these challenges, we propose Elastic Multi-Gradient Descent (EMGD), which formulates PCL as a dynamic multi-objective optimization problem. EMGD introduces task-specific elastic factors to guide gradient updates toward Pareto-optimal directions, ensuring balanced learning across tasks. Additionally, we develop a gradient-guided memory editing mechanism that aligns rehearsal data with the optimized descent direction, mitigating memory-induced interference. Theoretical analysis shows that accumulation points of the EMGD update rule are Pareto critical under the proposed formulation, and extensive experiments on image classification benchmarks demonstrate that EMGD significantly outperforms existing methods, including state-of-the-art PCL, MTL, and SCL approaches.
- Open Access
Taming Preconditioner Drift: Unlocking the Potential of Second-Order Optimizers for Federated Learning on Non-IID Data
Second-order optimizers can significantly accelerate large-scale training, yet their naive federated variants are often unstable or even diverge on non-IID data. We show that a key culprit is \emph{preconditioner drift}: client-side second-order training induces heterogeneous \emph{curvature-defined geometries} (i.e., preconditioner coordinate systems), and server-side model averaging updates computed under incompatible metrics, corrupting the global descent direction. To address this geometric mismatch, we propose \texttt{FedPAC}, a \emph{preconditioner alignment and correction} framework for reliable federated second-order optimization. \texttt{FedPAC} explicitly decouples parameter aggregation from geometry synchronization by: (i) \textbf{Alignment} (i.e.,aggregating local preconditioners into a global reference and warm-starting clients via global preconditioner); and (ii) \textbf{Correction} (i.e., steering local preconditioned updates using a global preconditioned direction to suppress long-term drift). We provide drift-coupled non-convex convergence guarantees with linear speedup under partial participation. Empirically, \texttt{FedPAC} consistently improves stability and accuracy across vision and language tasks, achieving up to $5.8\%$ absolute accuracy gain on CIFAR-100 with ViTs. Code is available at https://anonymous.4open.science/r/FedPAC-8B24.Open AccessTaming Preconditioner Drift: Unlocking the Potential of Second-Order Optimizers for Federated Learning on Non-IID Data
Second-order optimizers can significantly accelerate large-scale training, yet their naive federated variants are often unstable or even diverge on non-IID data. We show that a key culprit is \emph{preconditioner drift}: client-side second-order training induces heterogeneous \emph{curvature-defined geometries} (i.e., preconditioner coordinate systems), and server-side model averaging updates computed under incompatible metrics, corrupting the global descent direction. To address this geometric mismatch, we propose \texttt{FedPAC}, a \emph{preconditioner alignment and correction} framework for reliable federated second-order optimization. \texttt{FedPAC} explicitly decouples parameter aggregation from geometry synchronization by: (i) \textbf{Alignment} (i.e.,aggregating local preconditioners into a global reference and warm-starting clients via global preconditioner); and (ii) \textbf{Correction} (i.e., steering local preconditioned updates using a global preconditioned direction to suppress long-term drift). We provide drift-coupled non-convex convergence guarantees with linear speedup under partial participation. Empirically, \texttt{FedPAC} consistently improves stability and accuracy across vision and language tasks, achieving up to $5.8\%$ absolute accuracy gain on CIFAR-100 with ViTs. Code is available at https://anonymous.4open.science/r/FedPAC-8B24. - 11 December 2025
- Published by Elsevier in Pattern Recognition
- Vol. 174
Enhancing the impact of model performance gains for semi-supervised medical image segmentation
- 11 December 2025
- Published by Elsevier in Pattern Recognition
- Vol. 174
Enhancing the impact of model performance gains for semi-supervised medical image segmentation
- Open Access
- 6 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Gloss-Free Sign Language Translation With Optical-Flow Guided Two-Stream Network
- 6 December 2025
- Published by Association for Computing Machinery (ACM)
- p. 1-8
Open AccessGloss-Free Sign Language Translation With Optical-Flow Guided Two-Stream Network
- Open Access
ILoRA: Federated Learning with Low-Rank Adaptation for Heterogeneous Client Aggregation
Federated Learning with Low-Rank Adaptation (LoRA) faces three critical challenges under client heterogeneity: (1) Initialization-Induced Instability due to random initialization misaligning client subspaces; (2) Rank Incompatibility and Aggregation Error when averaging LoRA parameters of different ranks, which biases the global model; and (3) exacerbated Client Drift under Non-IID Data, impairing generalization. To address these challenges, we propose ILoRA, a unified framework that integrates three core innovations: a QR-based orthonormal initialization to ensure all clients start in a coherent subspace; a Concatenated QR Aggregation mechanism that fuses heterogeneous-rank updates via concatenation and decomposition, preserving information while maintaining dimension alignment; and an AdamW optimizer with rank-aware control variates to correct local updates and mitigate client drift. Supported by theoretical convergence guarantees, extensive experiments on vision and NLP benchmarks demonstrate that ILoRA consistently achieves superior accuracy and convergence stability compared to existing federated LoRA methods.
Open AccessILoRA: Federated Learning with Low-Rank Adaptation for Heterogeneous Client Aggregation
Federated Learning with Low-Rank Adaptation (LoRA) faces three critical challenges under client heterogeneity: (1) Initialization-Induced Instability due to random initialization misaligning client subspaces; (2) Rank Incompatibility and Aggregation Error when averaging LoRA parameters of different ranks, which biases the global model; and (3) exacerbated Client Drift under Non-IID Data, impairing generalization. To address these challenges, we propose ILoRA, a unified framework that integrates three core innovations: a QR-based orthonormal initialization to ensure all clients start in a coherent subspace; a Concatenated QR Aggregation mechanism that fuses heterogeneous-rank updates via concatenation and decomposition, preserving information while maintaining dimension alignment; and an AdamW optimizer with rank-aware control variates to correct local updates and mitigate client drift. Supported by theoretical convergence guarantees, extensive experiments on vision and NLP benchmarks demonstrate that ILoRA consistently achieves superior accuracy and convergence stability compared to existing federated LoRA methods.