arXiv subjects

Yong Xia

Publications and source records attributed to Yong Xia.

At least 37 records · Page 2Linked to original sources

arXiv2026

SegRap2025: A Benchmark of Gross Tumor Volume and Lymph Node Clinical Target Volume Segmentation for Radiotherapy Planning of Nasopharyngeal Carcinoma

Accurate delineation of Gross Tumor Volume (GTV), Lymph Node Clinical Target Volume (LN CTV), and Organ-at-Risk (OAR) from Computed Tomography (CT) scans is essential for precise radiotherapy planning in Nasopharyngeal Carcinoma (NPC). Building upon SegRap2023, which focused on OAR and GTV segmentation using single-center paired non-contrast CT (ncCT) and contrast-enhanced CT (ceCT) scans, the SegRap2025 challenge aims to enhance the generalizability and robustness of segmentation models across imaging centers and modalities. SegRap2025 comprises two tasks: Task01 addresses GTV segmentation using paired CT from the SegRap2023 dataset, with an additional external testing set to evaluate cross-center generalization, and Task02 focuses on LN CTV segmentation using multi-center training data and an unseen external testing set, where each case contains paired CT scans or a single modality, emphasizing both cross-center and cross-modality robustness. This paper presents the challenge setup and provides a comprehensive analysis of the solutions submitted by ten participating teams. For GTV segmentation task, the top-performing models achieved average Dice Similarity Coefficient (DSC) of 74.61% and 56.79% on the internal and external testing cohorts, respectively. For LN CTV segmentation task, the highest average DSC values reached 60.24%, 60.50%, and 57.23% on paired CT, ceCT-only, and ncCT-only subsets, respectively. SegRap2025 establishes a large-scale multi-center, multi-modality benchmark for evaluating the generalization and robustness in radiotherapy target segmentation, providing valuable insights toward clinically applicable automated radiotherapy planning systems. The benchmark is available at: https://hilab-git.github.io/SegRap2025_Challenge.

eess.IV↗

arXiv2026

V-Loop: Visual Logical Loop Verification for Hallucination Detection in Medical Visual Question Answering

Multimodal Large Language Models (MLLMs) have shown remarkable capability in assisting disease diagnosis in medical visual question answering (VQA). However, their outputs remain vulnerable to hallucinations (i.e., responses that contradict visual facts), posing significant risks in high-stakes medical scenarios. Recent introspective detection methods, particularly uncertainty-based approaches, offer computational efficiency but are fundamentally indirect, as they estimate predictive uncertainty for an image-question pair rather than verifying the factual correctness of a specific answer. To address this limitation, we propose Visual Logical Loop Verification (V-Loop), a training-free and plug-and-play framework for hallucination detection in medical VQA. V-Loop introduces a bidirectional reasoning process that forms a visually grounded logical loop to verify factual correctness. Given an input, the MLLM produces an answer for the primary input pair. V-Loop extracts semantic units from the primary QA pair, generates a verification question by conditioning on the answer unit to re-query the question unit, and enforces visual attention consistency to ensure answering both primary question and verification question rely on the same image evidence. If the verification answer matches the expected semantic content, the logical loop closes, indicating factual grounding; otherwise, the primary answer is flagged as hallucinated. Extensive experiments on multiple medical VQA benchmarks and MLLMs show that V-Loop consistently outperforms existing introspective methods, remains highly efficient, and further boosts uncertainty-based approaches when used in combination.

cs.CV↗

arXiv2026

PAINT: Pathology-Aware Integrated Next-Scale Transformation for Virtual Immunohistochemistry

Virtual immunohistochemistry (IHC) aims to computationally synthesize molecular staining patterns from routine Hematoxylin and Eosin (H\&E) images, offering a cost-effective and tissue-efficient alternative to traditional physical staining. However, this task is particularly challenging: H\&E morphology provides ambiguous cues about protein expression, and similar tissue structures may correspond to distinct molecular states. Most existing methods focus on direct appearance synthesis to implicitly achieve cross-modal generation, often resulting in semantic inconsistencies due to insufficient structural priors. In this paper, we propose Pathology-Aware Integrated Next-Scale Transformation (PAINT), a visual autoregressive framework that reformulates the synthesis process as a structure-first conditional generation task. Unlike direct image translation, PAINT enforces a causal order by resolving molecular details conditioned on a global structural layout. Central to this approach is the introduction of a Spatial Structural Start Map (3S-Map), which grounds the autoregressive initialization in observed morphology, ensuring deterministic, spatially aligned synthesis. Experiments on the IHC4BC and MIST datasets demonstrate that PAINT outperforms state-of-the-art methods in structural fidelity and clinical downstream tasks, validating the potential of structure-guided autoregressive modeling.

cs.CV↗

arXiv2026

Upper limits on microhertz gravitational waves from supermassive black-hole binaries using PSR J1909-3744 data from the second IPTA data release

We present the results of a search for gravitational waves (GWs) from individual sources using high-cadence observations of PSR J1909$-$3744 obtained during an intensive observing campaign with the International Pulsar Timing Array second data release (IPTA-DR2) between July 2010 and November 2012. The observations, conducted at three different radio frequencies with the Nançay Radio Telescope (NRT) and Parkes Telescope (PKS) and five frequencies with the Green Bank Telescope (GBT), enabled precise corrections for dispersion measure effects and scattering variations. After these corrections, the timing residuals showed an unmodeled periodic noise component with an amplitude of 340 ns. Our analysis yields upper limits on the GW strain from individual sources, constraining it to be below $1.9 \times 10^{-14}$ at 71 nHz and $2.3 \times 10^{-13}$ at 1 \textmu Hz for average sky locations, while for optimal source locations the limits improve to $6.2 \times 10^{-15}$ and $8.9 \times 10^{-14}$ at the same frequencies, respectively. Our new limits are about a factor of 1.52 more stringent than those of Perera et al. based on an earlier EPTA data.

astro-ph.HE↗

arXiv2025

Mini-Extragradient Methods

The Extragradient (EG) method stands as a cornerstone algorithm for solving monotone nonlinear equations but faces two important unresolved challenges: (i) how to select stepsizes without relying on the global Lipschitz constant or expensive line-search procedures, and (ii) how to reduce the two full evaluations of the mapping required per iteration to effectively one, without compromising convergence guarantees or computational efficiency. To address the first challenge, we propose the Greedy Mini-Extragradient (Mini-EG) method, which updates only the coordinate associated with the dominant component of the mapping at each extragradient step. This design capitalizes on componentwise Lipschitz constants that are far easier to estimate than the classical global Lipschitz constant. To further lower computational cost, we introduce a Random Mini-EG variant that replaces full mapping evaluations by sampling only a single coordinate per extragradient step. Although this resolves the second challenge from a theoretical standpoint, its practical efficiency remains limited. To bridge this gap, we develop the Watchdog-Max strategy, motivated by the slow decay of dominant component magnitudes. Instead of evaluating the full mapping, Watchdog-Max identifies and tracks only two coordinates at each extragradient step, dramatically reducing per-iteration cost while retaining strong practical performance. We establish convergence guarantees and rate analyses for all proposed methods. In particular, Greedy Mini-EG achieves enhanced convergence rates that surpass the classical guarantees of the vanilla EG method in several standard application settings. Numerical experiments on regularized decentralized logistic regression and compressed sensing show speedups exceeding $13\times$ compared with the classical EG method on both synthetic and real datasets.

math.OC↗

arXiv2025

MIRROR: Multi-Modal Pathological Self-Supervised Representation Learning via Modality Alignment and Retention

Histopathology and transcriptomics are fundamental modalities in oncology, encapsulating the morphological and molecular aspects of the disease. Multi-modal self-supervised learning has demonstrated remarkable potential in learning pathological representations by integrating diverse data sources. Conventional multi-modal integration methods primarily emphasize modality alignment, while paying insufficient attention to retaining the modality-specific structures. However, unlike conventional scenarios where multi-modal inputs share highly overlapping features, histopathology and transcriptomics exhibit pronounced heterogeneity, offering orthogonal yet complementary insights. Histopathology provides morphological and spatial context, elucidating tissue architecture and cellular topology, whereas transcriptomics delineates molecular signatures through gene expression patterns. This inherent disparity introduces a major challenge in aligning them while maintaining modality-specific fidelity. To address these challenges, we present MIRROR, a novel multi-modal representation learning method designed to foster both modality alignment and retention. MIRROR employs dedicated encoders to extract comprehensive features for each modality, which is further complemented by a modality alignment module to achieve seamless integration between phenotype patterns and molecular profiles. Furthermore, a modality retention module safeguards unique attributes from each modality, while a style clustering module mitigates redundancy and enhances disease-relevant information by modeling and aligning consistent pathological signatures within a clustering space. Extensive evaluations on TCGA cohorts for cancer subtyping and survival analysis highlight MIRROR's superior performance, demonstrating its effectiveness in constructing comprehensive oncological feature representations and benefiting the cancer diagnosis.

cs.CV↗

arXiv2025

From Generality to Specificity: Prior-Driven Optimal Sparse Transformation in Compressed Sensing

This paper introduces a new paradigm for sparse transformation: the Prior-to-Posterior Sparse Transform (POST) framework, designed to overcome long-standing limitation on generalization and specificity in classical sparse transforms for compressed sensing. POST systematically unifies the generalization capacity of any existing transform domains with the specificity of reference knowledge, enabling flexible adaptation to diverse signal characteristics. Within this framework, we derive an explicit sparse transform domain termed HOT, which adaptively handles both real and complex-valued signals. We theoretically establish HOT's sparse representation properties under single and multiple reference settings, demonstrating its ability to preserve generalization while enhancing specificity even under weak reference information. Extensive experiments confirm that HOT delivers substantial meta-gains across audio sensing, 5G channel estimation, and image compression tasks, consistently boosting multiple compressed sensing algorithms under diverse multimodal settings with negligible computational overhead.

math.OC↗

arXiv2025

Accelerated Dinkelbach method

The classical Dinkelbach method (1967) solves fractional programming via a parametric approach, generating a decreasing upper bound sequence that converges to the optimum. Its important variant, the interval Dinkelbach method (1991), constructs convergent upper and lower bound sequences that bracket the solution and achieve quadratic and superlinear convergence, respectively, under the assumption that the parametric function is twice continuously differentiable. However, this paper demonstrates that a minimal correction, applied solely to the upper bound iterate, is sufficient to boost the convergence of the method, achieving superquadratic and cubic rates for the upper and lower bound sequences, respectively. By strategically integrating this correction, we develop a globally convergent, non-monotone, and accelerated Dinkelbach algorithm-the first of its kind, to our knowledge. Under sufficient differentiability, the new method achieves an asymptotic average convergence order of at least the square root of 5 per iteration, surpassing the quadratic order of the original algorithm. Crucially, this acceleration is achieved while maintaining the key practicality of solving only a single subproblem per iteration.

math.OC↗

arXiv2025

ECG-aBcDe: Overcoming Model Dependence, Encoding ECG into a Universal Language for Any LLM

Large Language Models (LLMs) hold significant promise for electrocardiogram (ECG) analysis, yet challenges remain regarding transferability, time-scale information learning, and interpretability. Current methods suffer from model-specific ECG encoders, hindering transfer across LLMs. Furthermore, LLMs struggle to capture crucial time-scale information inherent in ECGs due to Transformer limitations. And their black-box nature limits clinical adoption. To address these limitations, we introduce ECG-aBcDe, a novel ECG encoding method that transforms ECG signals into a universal ECG language readily interpretable by any LLM. By constructing a hybrid dataset of ECG language and natural language, ECG-aBcDe enables direct fine-tuning of pre-trained LLMs without architectural modifications, achieving "construct once, use anywhere" capability. Moreover, the bidirectional convertibility between ECG and ECG language of ECG-aBcDe allows for extracting attention heatmaps from ECG signals, significantly enhancing interpretability. Finally, ECG-aBcDe explicitly represents time-scale information, mitigating Transformer limitations. This work presents a new paradigm for integrating ECG analysis with LLMs. Compared with existing methods, our method achieves competitive performance on ROUGE-L and METEOR. Notably, it delivers significant improvements in the BLEU-4, with improvements of 2.8 times and 3.9 times in in-dataset and cross-dataset evaluations, respectively, reaching scores of 42.58 and 30.76. These results provide strong evidence for the feasibility of the new paradigm.

cs.AI↗

arXiv2025

Lipschitz-free Projected Subgradient Method with Time-varying Step-size

We introduce a novel family of time-varying step-sizes for the classical projected subgradient method, offering optimal ergodic convergence. Importantly, this approach does not depend on the Lipschitz assumption of the objective function, thereby broadening the convergence result of projected subgradient method to non-Lipschitz case.

math.OC↗

arXiv2025

Unified Start, Personalized End: Progressive Pruning for Efficient 3D Medical Image Segmentation

3D medical image segmentation often faces heavy resource and time consumption, limiting its scalability and rapid deployment in clinical environments. Existing efficient segmentation models are typically static and manually designed prior to training, which restricts their adaptability across diverse tasks and makes it difficult to balance performance with resource efficiency. In this paper, we propose PSP-Seg, a progressive pruning framework that enables dynamic and efficient 3D segmentation. PSP-Seg begins with a redundant model and iteratively prunes redundant modules through a combination of block-wise pruning and a functional decoupling loss. We evaluate PSP-Seg on five public datasets, benchmarking it against seven state-of-the-art models and six efficient segmentation models. Results demonstrate that the lightweight variant, PSP-Seg-S, achieves performance on par with nnU-Net while reducing GPU memory usage by 42-45%, training time by 29-48%, and parameter number by 83-87% across all datasets. These findings underscore PSP-Seg's potential as a cost-effective yet high-performing alternative for widespread clinical application.

cs.CV↗

arXiv2025

MedSeqFT: Sequential Fine-tuning Foundation Models for 3D Medical Image Segmentation

Foundation models have become a promising paradigm for advancing medical image analysis, particularly for segmentation tasks where downstream applications often emerge sequentially. Existing fine-tuning strategies, however, remain limited: parallel fine-tuning isolates tasks and fails to exploit shared knowledge, while multi-task fine-tuning requires simultaneous access to all datasets and struggles with incremental task integration. To address these challenges, we propose MedSeqFT, a sequential fine-tuning framework that progressively adapts pre-trained models to new tasks while refining their representational capacity. MedSeqFT introduces two core components: (1) Maximum Data Similarity (MDS) selection, which identifies downstream samples most representative of the original pre-training distribution to preserve general knowledge, and (2) Knowledge and Generalization Retention Fine-Tuning (K&G RFT), a LoRA-based knowledge distillation scheme that balances task-specific adaptation with the retention of pre-trained knowledge. Extensive experiments on two multi-task datasets covering ten 3D segmentation tasks demonstrate that MedSeqFT consistently outperforms state-of-the-art fine-tuning strategies, yielding substantial performance gains (e.g., an average Dice improvement of 3.0%). Furthermore, evaluations on two unseen tasks (COVID-19-20 and Kidney) verify that MedSeqFT enhances transferability, particularly for tumor segmentation. Visual analyses of loss landscapes and parameter variations further highlight the robustness of MedSeqFT. These results establish sequential fine-tuning as an effective, knowledge-retentive paradigm for adapting foundation models to evolving clinical tasks. Code will be released.

cs.CV↗

arXiv2025

Toward Robust Medical Fairness: Debiased Dual-Modal Alignment via Text-Guided Attribute-Disentangled Prompt Learning for Vision-Language Models

Ensuring fairness across demographic groups in medical diagnosis is essential for equitable healthcare, particularly under distribution shifts caused by variations in imaging equipment and clinical practice. Vision-language models (VLMs) exhibit strong generalization, and text prompts encode identity attributes, enabling explicit identification and removal of sensitive directions. However, existing debiasing approaches typically address vision and text modalities independently, leaving residual cross-modal misalignment and fairness gaps. To address this challenge, we propose DualFairVL, a multimodal prompt-learning framework that jointly debiases and aligns cross-modal representations. DualFairVL employs a parallel dual-branch architecture that separates sensitive and target attributes, enabling disentangled yet aligned representations across modalities. Approximately orthogonal text anchors are constructed via linear projections, guiding cross-attention mechanisms to produce fused features. A hypernetwork further disentangles attribute-related information and generates instance-aware visual prompts, which encode dual-modal cues for fairness and robustness. Prototype-based regularization is applied in the visual branch to enforce separation of sensitive features and strengthen alignment with textual anchors. Extensive experiments on eight medical imaging datasets across four modalities show that DualFairVL achieves state-of-the-art fairness and accuracy under both in- and out-of-distribution settings, outperforming full fine-tuning and parameter-efficient baselines with only 3.6M trainable parameters. Code will be released upon publication.

cs.CV↗

arXiv2025

Best Subset Selection: Optimal Pursuit for Feature Selection and Elimination

This paper introduces two novel criteria: one for feature selection and another for feature elimination in the context of best subset selection, which is a benchmark problem in statistics and machine learning. From the perspective of optimization, we revisit the classical selection and elimination criteria in traditional best subset selection algorithms, revealing that these classical criteria capture only partial variations of the objective function after the entry or exit of features. By formulating and solving optimization subproblems for feature entry and exit exactly, new selection and elimination criteria are proposed, proved as the optimal decisions for the current entry-and-exit process compared to classical criteria. Replacing the classical selection and elimination criteria with the proposed ones generates a series of enhanced best subset selection algorithms. These generated algorithms not only preserve the theoretical properties of the original algorithms but also achieve significant meta-gains without increasing computational cost across various scenarios and evaluation metrics on multiple tasks such as compressed sensing and sparse regression.

math.OC↗

arXiv2025

From Few to More: Scribble-based Medical Image Segmentation via Masked Context Modeling and Continuous Pseudo Labels

Scribble-based weakly supervised segmentation methods have shown promising results in medical image segmentation, significantly reducing annotation costs. However, existing approaches often rely on auxiliary tasks to enforce semantic consistency and use hard pseudo labels for supervision, overlooking the unique challenges faced by models trained with sparse annotations. These models must predict pixel-wise segmentation maps from limited data, making it crucial to handle varying levels of annotation richness effectively. In this paper, we propose MaCo, a weakly supervised model designed for medical image segmentation, based on the principle of "from few to more." MaCo leverages Masked Context Modeling (MCM) and Continuous Pseudo Labels (CPL). MCM employs an attention-based masking strategy to perturb the input image, ensuring that the model's predictions align with those of the original image. CPL converts scribble annotations into continuous pixel-wise labels by applying an exponential decay function to distance maps, producing confidence maps that represent the likelihood of each pixel belonging to a specific category, rather than relying on hard pseudo labels. We evaluate MaCo on three public datasets, comparing it with other weakly supervised methods. Our results show that MaCo outperforms competing methods across all datasets, establishing a new record in weakly supervised medical image segmentation.

cs.CV↗

arXiv2025

Convergence Analysis of Reshaped Wirtinger Flow with Random Initialization for Phase Retrieval

This paper investigates phase retrieval using the Reshaped Wirtinger Flow (RWF) algorithm, focusing on recovering target vector $\vx \in \R^n$ from magnitude measurements $y_i = \left| \langle \va_i, \vx \rangle \right|, \; i = 1, \ldots, m,$ under random initialization, where $\va_i \in \R^n$ are measurement vectors. For Gaussian measurement designs, we prove that when $m\ge O(n \log^2 n\log^3 m)$, the RWF algorithm with random initialization achieves $ε$-accuracy within $O\big(\log n + \log(1/ε)\big)$ iterations, thereby attaining nearly optimal sample and computational complexities comparable to those previously established for spectrally initialized methods. Numerical experiments demonstrate that the convergence rate is robust to initialization randomness and remains stable even with larger step sizes.

math.OC↗

arXiv2025

Cycle Context Verification for In-Context Medical Image Segmentation

In-context learning (ICL) is emerging as a promising technique for achieving universal medical image segmentation, where a variety of objects of interest across imaging modalities can be segmented using a single model. Nevertheless, its performance is highly sensitive to the alignment between the query image and in-context image-mask pairs. In a clinical scenario, the scarcity of annotated medical images makes it challenging to select optimal in-context pairs, and fine-tuning foundation ICL models on contextual data is infeasible due to computational costs and the risk of catastrophic forgetting. To address this challenge, we propose Cycle Context Verification (CCV), a novel framework that enhances ICL-based medical image segmentation by enabling self-verification of predictions and accordingly enhancing contextual alignment. Specifically, CCV employs a cyclic pipeline in which the model initially generates a segmentation mask for the query image. Subsequently, the roles of the query and an in-context pair are swapped, allowing the model to validate its prediction by predicting the mask of the original in-context image. The accuracy of this secondary prediction serves as an implicit measure of the initial query segmentation. A query-specific prompt is introduced to alter the query image and updated to improve the measure, thereby enhancing the alignment between the query and in-context pairs. We evaluated CCV on seven medical image segmentation datasets using two ICL foundation models, demonstrating its superiority over existing methods. Our results highlight CCV's ability to enhance ICL-based segmentation, making it a robust solution for universal medical image segmentation. The code will be available at https://github.com/ShishuaiHu/CCV.

cs.CV↗

arXiv2025

Split-Merge Revisited: A Scalable Approach to Generalized Eigenvalue Problems

The generalized eigenvalue problem (GEP) serves as a cornerstone in a wide range of applications in numerical linear algebra and scientific computing. However, traditional approaches that aim to maximize the classical Rayleigh quotient often suffer from numerical instability and limited computational efficiency, especially in large-scale settings. In this work, we explore an alternative difference-based formulation of GEP by minimizing a structured quadratic polynomial objective, which enables the application of efficient first-order optimization methods. We establish global convergence guarantees for these methods without requiring line search, and further introduce a transform-domain perspective that reveals the intrinsic connection and performance gap between classical first-order algorithms and the power method. Based on this insight, we develop an accelerated preconditioned mirror descent algorithm, which allows for flexible preconditioner design and improved convergence behavior. Lastly, we extend the recently proposed Split-Merge algorithm to the general GEP setting, incorporating richer second-order information to further accelerate convergence. Empirical results on both synthetic and real-world datasets demonstrate that our proposed methods achieve significant improvements over existing baselines in terms of both computational efficiency and numerical stability.

math.OC↗