arXiv subjects

Yehui Yang

Publications and source records attributed to Yehui Yang.

At least 19 recordsLinked to original sources

arXiv2026

PhysAgent: A Multi-Agent Framework for Reliable Remote Heart Rate Estimation

Remote photoplethysmography (rPPG) enables non-contact heart-rate estimation from facial videos, but its weak physiological signal is easily corrupted by motion, illumination changes, occlusion, skin-appearance variation, and device noise. Existing rPPG methods typically rely on a single model to directly predict heart rate or recover pulse waveforms, while different strong estimators may produce conflicting yet individually plausible candidates for the same video. To resolve these conflicts, we propose PhysAgent, an inference-time multi-agent candidate-verification framework. Unlike direct prediction approaches, PhysAgent neither trains a new base rPPG model nor asks Multimodal Large Language Models (MLLMs) to output heart rate directly. In contrast, it treats outputs from multiple base estimators as physiological hypotheses to be verified and uses a lightweight 4B MLLM, Qwen3-VL-4B, to drive multi-agent reasoning over video conditions, signal reliability, and candidate disagreement. A deterministic physiological verifier checks the fusion proposal, and a reproducible numerical fusion process produces the final heart rate. Experimental results on multiple public rPPG benchmarks show that PhysAgent improves fusion stability and reliability across different datasets and source-domain settings, while avoiding the irreproducibility and physiological inconsistency of direct MLLM prediction or unconstrained ensemble fusion. The code will be released soon.

cs.CV↗

arXiv2026

CriticGen: Generation-Aware Evaluation as Actionable Feedback

Current evaluation methods for large language models are coarse-grained and decoupled from generation, producing generic explanations that fail to provide actionable feedback for model improvement. We propose CriticGen, a fine-grained, generation-aware evaluation framework that turns evaluation into actionable control for answer improvement. CriticGen first generates sample-specific evaluation dimensions and scoring criteria under high-level categories such as subjective, objective, and self-derived constraints. These criteria then serve as a dynamic rubric for jointly producing a score, a reason, an executable refinement suggestion, and a refined answer. This rubric-conditioned refinement process enables models to diagnose flaws and perform targeted answer improvement. Experimental results show that fine-grained evaluation should be both instance-specific and actionable. CriticGen induces higher-quality rubrics, improving relevance/coverage from 3.33/4.03 to 3.97/4.24. CriticGen also achieves the best score correlations, with 0.9556 Pearson and 0.9560 Spearman, and raises the F1 of criterion-grounded reasons and executable suggestions from 0.6369/0.5994 to 0.7554/0.7900. Crucially, its feedback translates into reliable answer improvement, improving 73.17% of answers with a 93.28% non-degradation rate.

cs.AI↗

arXiv2026

Hypothesis-Driven Skill Optimization for LLM Agents

External skills can improve action-oriented LLM agents without changing model weights, but persistent skill updates are risky when they are distilled from sparse or noisy trajectories. A plausible reflection may encode a useful procedure, a spurious shortcut, or a rule that the target executor cannot reliably follow. We propose Hypothesis-Driven Skill Optimization (HDSO), a train-free framework in which both the skill curator and the agent executor are frozen inference endpoints. The curator observes executor traces, proposes a falsifiable hypothesis with an explicit validation plan, instantiates the hypothesis as a candidate skill package, validates the package through paired control/treatment executions, reviews behavior differences, and consolidates only supported candidates into an approved repository. The executor consumes approved skills through progressive disclosure, preserving the executor-only path when no skill is selected. On ALFWorld, HDSO improves executor-only baselines by +6.9 Avg. SR points for Qwen3-8B and +4.0 points for Qwen3.6-27B. Under 20% randomly flipped success/failure feedback during skill discovery and validation, HDSO preserves a +7.1-point gain for Qwen3-8B. Transfer and heterogeneous-pair diagnostics further show that validated repositories can be useful beyond the run that produced them, but cross-model curation succeeds only when curator diagnosis, executor capability, and validation evidence align. HDSO provides an auditable skill lifecycle for frozen action agents rather than an unconstrained memory accumulation procedure.

cs.AI↗

arXiv2026

FCMBench-Video: Benchmarking Document Video Intelligence

Document understanding is a critical capability in financial credit review, onboarding, and remote verification, where both decision accuracy and evidence traceability matter. Compared with static document images, document videos present a temporally redundant and sequentially unfolding evidence stream, require evidence integration across frames, and preserve acquisition-process cues relevant to authenticity-sensitive and anti-fraud review. We introduce FCMBench-Video, a benchmark for document-video intelligence that evaluates document perception, temporal grounding, and evidence-grounded reasoning under realistic capture conditions. For privacy-compliant yet realistic data at scale, we organize construction as an atomic-acquisition and composition workflow that records reusable single-document clips, applies controlled degradations, and assembles long-form multi-document videos with prescribed temporal spans. FCMBench-Video is built from 495 atomic videos composed into 1,200 long-form videos paired with 11,322 expert-annotated question--answer instances, covering 28 document types over 20s--60s duration tiers and 5,960 Chinese / 5,362 English instances. Evaluations on nine recent Video-MLLMs show that FCMBench-Video provides meaningful separation across systems and capabilities: counting is the most duration-sensitive task, Cross-Document Validation and Evidence-Grounded Selection probe higher-level evidence integration, and Visual Prompt Injection provides a complementary robustness dimension. The overall score distribution is broad and approximately bell-shaped, indicating a benchmark that is neither saturated nor dominated by trivial cases. Together, these results position FCMBench-Video as a reproducible benchmark for tracking Video-MLLM progress on document-video understanding and probing capability boundaries in authenticity-sensitive credit-domain applications.

cs.CV↗

arXiv2026

MAT-Cell: A Multi-Agent Tree-Structured Reasoning Framework for Batch-Level Single-Cell Annotation

Automated single-cell annotation is difficult when the most abundant genes are not the most discriminative ones, or when a target state is poorly covered by a fixed reference atlas. GPTCelltype-style one-shot prompting allows large language models (LLMs) to produce plausible labels from generic expression signals, while reference-based annotators can force unfamiliar states into the nearest known category. We propose MAT-Cell, a prompt-driven framework for batch-level single-cell annotation that separates evidence grounding from label decision. MAT-Cell first uses Reverse Verification Query (RVQ) to combine tissue context, observed differentially expressed genes, and LLM-elicited biological priors into structured candidate-specific premises. Verifier agents then convert these premises into explicit premise-to-claim reasoning trees, and bounded multi-round debate compares,challenges, and revises the resulting claims before consensus or final adjudication.The returned Syllogistic Derivation Tree (SDT) provides an auditable debate trace rather than a formal proof of the annotation. In open-candidate benchmarks across five datasets, a locally deployed Qwen3-30B model with MAT-Cell achieves 75.5% average accuracy, compared with 64.2% for the strongest evaluated CoT baseline and 51.9% for the strongest evaluated scPilot variant. In oracle-candidate bench-marks across three species,MAT-Cell remains competitive across backbones, and local inference substantially reduces monetary cost for batch annotation. Code is available at: https://anonymous.4open.science/r/MATCell-4067

q-bio.QM↗

arXiv2026

Revisiting Cross-Attention Mechanisms: Leveraging Beneficial Noise for Domain-Adaptive Learning

Unsupervised Domain Adaptation (UDA) seeks to transfer knowledge from a labeled source domain to an unlabeled target domain but often suffers from severe domain and scale gaps that degrade performance. Existing cross-attention-based transformers can align features across domains, yet they struggle to preserve content semantics under large appearance and scale variations. To explicitly address these challenges, we introduce the concept of beneficial noise, which regularizes cross-attention by injecting controlled perturbations, encouraging the model to ignore style distractions and focus on content. We propose the Domain-Adaptive Cross-Scale Matching (DACSM) framework, which consists of a Domain-Adaptive Transformer (DAT) for disentangling domain-shared content from domain-specific style, and a Cross-Scale Matching (CSM) module that adaptively aligns features across multiple resolutions. DAT incorporates beneficial noise into cross-attention, enabling progressive domain translation with enhanced robustness, yielding content-consistent and style-invariant representations. Meanwhile, CSM ensures semantic consistency under scale changes. Extensive experiments on VisDA-2017, Office-Home, and DomainNet demonstrate that DACSM achieves state-of-the-art performance, with up to +2.3% improvement over CDTrans on VisDA-2017. Notably, DACSM achieves a +5.9% gain on the challenging "truck" class of VisDA, evidencing the strength of beneficial noise in handling scale discrepancies. These results highlight the effectiveness of combining domain translation, beneficial-noise-enhanced attention, and scale-aware alignment for robust cross-domain representation learning.

cs.CV↗

arXiv2026

FCMBench: The First Large-scale Financial Credit Multimodal Benchmark for Real-world Applications

FCMBench is the first large-scale and privacy-compliant multimodal benchmark for real-world financial credit applications, covering tasks and robustness challenges from domain specific workflows and constraints. The current version of FCMBench covers 26 certificate types, with 5198 privacy-compliant images and 13806 paired VQA samples. It evaluates models on Perception and Reasoning tasks under real-world Robustness interferences, including 3 foundational perception tasks, 4 credit-specific reasoning tasks demanding decision-oriented visual evidence interpretation, and 10 real-world challenges for rigorous robustness stress testing. Moreover, FCMBench offers privacy-compliant realism with minimal leakage risk through in-house scenario-aware captures of manually synthesized templates, without any publicly released images. We conduct extensive evaluations of 28 state-of-the-art vision-language models spanning 14 AI companies and research institutes. Among them, Gemini 3 Pro achieves the best F1 score as a commercial model (65.16), Kimi-K2.5 achieves the best score as an open-source baseline (60.58). The mean and the std. of all tested models is 44.8 and 10.3 respectively, indicating that FCMBench is non-trivial and provides strong resolution for separating modern vision-language model capabilities. Robustness evaluations reveal that even top-performing models experience notable performance degradation under the designed challenges. We have open-sourced this benchmark to advance AI research in the credit domain and provide a domain-specific task for real-world AI applications.

cs.CV↗

arXiv2025

DATE: Dynamic Absolute Time Enhancement for Long Video Understanding

Long video understanding remains a fundamental challenge for multimodal large language models (MLLMs), particularly in tasks requiring precise temporal reasoning and event localization. Existing approaches typically adopt uniform frame sampling and rely on implicit position encodings to model temporal order. However, these methods struggle with long-range dependencies, leading to critical information loss and degraded temporal comprehension. In this paper, we propose Dynamic Absolute Time Enhancement (DATE) that enhances temporal awareness in MLLMs through the Timestamp Injection Mechanism (TIM) and a semantically guided Temporal-Aware Similarity Sampling (TASS) strategy. Specifically, we interleave video frame embeddings with textual timestamp tokens to construct a continuous temporal reference system. We further reformulate the video sampling problem as a vision-language retrieval task and introduce a two-stage algorithm to ensure both semantic relevance and temporal coverage: enriching each query into a descriptive caption to better align with the vision feature, and sampling key event with a similarity-driven temporally regularized greedy strategy. Our method achieves remarkable improvements w.r.t. absolute time understanding and key event localization, resulting in state-of-the-art performance among 7B and 72B models on hour-long video benchmarks. Particularly, our 7B model even exceeds many 72B models on some benchmarks.

cs.CV↗

arXiv2024

Towards a Multimodal Large Language Model with Pixel-Level Insight for Biomedicine

In recent years, Multimodal Large Language Models (MLLM) have achieved notable advancements, demonstrating the feasibility of developing an intelligent biomedical assistant. However, current biomedical MLLMs predominantly focus on image-level understanding and restrict interactions to textual commands, thus limiting their capability boundaries and the flexibility of usage. In this paper, we introduce a novel end-to-end multimodal large language model for the biomedical domain, named MedPLIB, which possesses pixel-level understanding. Excitingly, it supports visual question answering (VQA), arbitrary pixel-level prompts (points, bounding boxes, and free-form shapes), and pixel-level grounding. We propose a novel Mixture-of-Experts (MoE) multi-stage training strategy, which divides MoE into separate training phases for a visual-language expert model and a pixel-grounding expert model, followed by fine-tuning using MoE. This strategy effectively coordinates multitask learning while maintaining the computational cost at inference equivalent to that of a single expert model. To advance the research of biomedical MLLMs, we introduce the Medical Complex Vision Question Answering Dataset (MeCoVQA), which comprises an array of 8 modalities for complex medical imaging question answering and image region understanding. Experimental results indicate that MedPLIB has achieved state-of-the-art outcomes across multiple medical visual language tasks. More importantly, in zero-shot evaluations for the pixel grounding task, MedPLIB leads the best small and large models by margins of 19.7 and 15.6 respectively on the mDice metric. The codes, data, and model checkpoints will be made publicly available at https://github.com/ShawnHuang497/MedPLIB.

cs.CV↗

arXiv2024

A Refer-and-Ground Multimodal Large Language Model for Biomedicine

With the rapid development of multimodal large language models (MLLMs), especially their capabilities in visual chat through refer and ground functionalities, their significance is increasingly recognized. However, the biomedical field currently exhibits a substantial gap in this area, primarily due to the absence of a dedicated refer and ground dataset for biomedical images. To address this challenge, we devised the Med-GRIT-270k dataset. It comprises 270k question-and-answer pairs and spans eight distinct medical imaging modalities. Most importantly, it is the first dedicated to the biomedical domain and integrating refer and ground conversations. The key idea is to sample large-scale biomedical image-mask pairs from medical segmentation datasets and generate instruction datasets from text using chatGPT. Additionally, we introduce a Refer-and-Ground Multimodal Large Language Model for Biomedicine (BiRD) by using this dataset and multi-task instruction learning. Extensive experiments have corroborated the efficacy of the Med-GRIT-270k dataset and the multi-modal, fine-grained interactive capabilities of the BiRD model. This holds significant reference value for the exploration and development of intelligent biomedical assistants.

cs.CV↗

arXiv2024

SegICL: A Multimodal In-context Learning Framework for Enhanced Segmentation in Medical Imaging

In the field of medical image segmentation, tackling Out-of-Distribution (OOD) segmentation tasks in a cost-effective manner remains a significant challenge. Universal segmentation models is a solution, which aim to generalize across the diverse modality of medical images, yet their effectiveness often diminishes when applied to OOD data modalities and tasks, requiring intricate fine-tuning of model for optimal performance. Few-shot learning segmentation methods are typically designed for specific modalities of data and cannot be directly transferred for use with another modality. Therefore, we introduce SegICL, a novel approach leveraging In-Context Learning (ICL) for image segmentation. Unlike existing methods, SegICL has the capability to employ text-guided segmentation and conduct in-context learning with a small set of image-mask pairs, eliminating the need for training the model from scratch or fine-tuning for OOD tasks (including OOD modality and dataset). Extensive experimental demonstrates a positive correlation between the number of shots and segmentation performance on OOD tasks. The performance of segmentation when provided thre-shots is approximately 1.5 times better than the performance in a zero-shot setting. This indicates that SegICL effectively address new segmentation tasks based on contextual information. Additionally, SegICL also exhibits comparable performance to mainstream models on OOD and in-distribution tasks. Our code will be released after paper review.

cs.CV↗

arXiv2024

FastqZip: An Improved Reference-Based Genome Sequence Lossy Compression Framework

Storing and archiving data produced by next-generation sequencing (NGS) is a huge burden for research institutions. Reference-based compression algorithms are effective in dealing with these data. Our work focuses on compressing FASTQ format files with an improved reference-based compression algorithm to achieve a higher compression ratio than other state-of-the-art algorithms. We propose FastqZip, which uses a new method mapping the sequence to reference for compression, allows reads-reordering and lossy quality scores, and the BSC or ZPAQ algorithm to perform final lossless compression for a higher compression ratio and relatively fast speed. Our method ensures the sequence can be losslessly reconstructed while allowing lossless or lossy compression for the quality scores. We reordered the reads to get a higher compression ratio. We evaluate our algorithms on five datasets and show that FastqZip can outperform the SOTA algorithm Genozip by around 10% in terms of compression ratio while having an acceptable slowdown.

cs.IT↗

arXiv2024

An Embeddable Implicit IUVD Representation for Part-based 3D Human Surface Reconstruction

To reconstruct a 3D human surface from a single image, it is crucial to simultaneously consider human pose, shape, and clothing details. Recent approaches have combined parametric body models (such as SMPL), which capture body pose and shape priors, with neural implicit functions that flexibly learn clothing details. However, this combined representation introduces additional computation, e.g. signed distance calculation in 3D body feature extraction, leading to redundancy in the implicit query-and-infer process and failing to preserve the underlying body shape prior. To address these issues, we propose a novel IUVD-Feedback representation, consisting of an IUVD occupancy function and a feedback query algorithm. This representation replaces the time-consuming signed distance calculation with a simple linear transformation in the IUVD space, leveraging the SMPL UV maps. Additionally, it reduces redundant query points through a feedback mechanism, leading to more reasonable 3D body features and more effective query points, thereby preserving the parametric body prior. Moreover, the IUVD-Feedback representation can be embedded into any existing implicit human reconstruction pipeline without requiring modifications to the trained neural networks. Experiments on the THuman2.0 dataset demonstrate that the proposed IUVD-Feedback representation improves the robustness of results and achieves three times faster acceleration in the query-and-infer process. Furthermore, this representation holds potential for generative applications by leveraging its inherent semantic information from the parametric body model.

cs.CV↗

arXiv2023

SynFundus-1M: A High-quality Million-scale Synthetic fundus images Dataset with Fifteen Types of Annotation

Large-scale public datasets with high-quality annotations are rarely available for intelligent medical imaging research, due to data privacy concerns and the cost of annotations. In this paper, we release SynFundus-1M, a high-quality synthetic dataset containing over one million fundus images in terms of \textbf{eleven disease types}. Furthermore, we deliberately assign four readability labels to the key regions of the fundus images. To the best of our knowledge, SynFundus-1M is currently the largest fundus dataset with the most sophisticated annotations. Leveraging over 1.3 million private authentic fundus images from various scenarios, we trained a powerful Denoising Diffusion Probabilistic Model, named SynFundus-Generator. The released SynFundus-1M are generated by SynFundus-Generator under predefined conditions. To demonstrate the value of SynFundus-1M, extensive experiments are designed in terms of the following aspect: 1) Authenticity of the images: we randomly blend the synthetic images with authentic fundus images, and find that experienced annotators can hardly distinguish the synthetic images from authentic ones. Moreover, we show that the disease-related vision features (e.g. lesions) are well simulated in the synthetic images. 2) Effectiveness for down-stream fine-tuning and pretraining: we demonstrate that retinal disease diagnosis models of either convolutional neural networks (CNN) or Vision Transformer (ViT) architectures can benefit from SynFundus-1M, and compared to the datasets commonly used for pretraining, models trained on SynFundus-1M not only achieve superior performance but also demonstrate faster convergence on various downstream tasks. SynFundus-1M is already public available for the open-source community.

cs.CV↗

arXiv2022

Multi-rater Prism: Learning self-calibrated medical image segmentation from multiple raters

In medical image segmentation, it is often necessary to collect opinions from multiple experts to make the final decision. This clinical routine helps to mitigate individual bias. But when data is multiply annotated, standard deep learning models are often not applicable. In this paper, we propose a novel neural network framework, called Multi-Rater Prism (MrPrism) to learn the medical image segmentation from multiple labels. Inspired by the iterative half-quadratic optimization, the proposed MrPrism will combine the multi-rater confidences assignment task and calibrated segmentation task in a recurrent manner. In this recurrent process, MrPrism can learn inter-observer variability taking into account the image semantic properties, and finally converges to a self-calibrated segmentation result reflecting the inter-observer agreement. Specifically, we propose Converging Prism (ConP) and Diverging Prism (DivP) to process the two tasks iteratively. ConP learns calibrated segmentation based on the multi-rater confidence maps estimated by DivP. DivP generates multi-rater confidence maps based on the segmentation masks estimated by ConP. The experimental results show that by recurrently running ConP and DivP, the two tasks can achieve mutual improvement. The final converged segmentation result of MrPrism outperforms state-of-the-art (SOTA) strategies on a wide range of medical image segmentation tasks.

eess.IV↗

arXiv2022

ExpNet: A unified network for Expert-Level Classification

Different from the general visual classification, some classification tasks are more challenging as they need the professional categories of the images. In the paper, we call them expert-level classification. Previous fine-grained vision classification (FGVC) has made many efforts on some of its specific sub-tasks. However, they are difficult to expand to the general cases which rely on the comprehensive analysis of part-global correlation and the hierarchical features interaction. In this paper, we propose Expert Network (ExpNet) to address the unique challenges of expert-level classification through a unified network. In ExpNet, we hierarchically decouple the part and context features and individually process them using a novel attentive mechanism, called Gaze-Shift. In each stage, Gaze-Shift produces a focal-part feature for the subsequent abstraction and memorizes a context-related embedding. Then we fuse the final focal embedding with all memorized context-related embedding to make the prediction. Such an architecture realizes the dual-track processing of partial and global information and hierarchical feature interactions. We conduct the experiments over three representative expert-level classification tasks: FGVC, disease classification, and artwork attributes classification. In these experiments, superior performance of our ExpNet is observed comparing to the state-of-the-arts in a wide range of fields, indicating the effectiveness and generalization of our ExpNet. The code will be made publicly available.

cs.CV↗

arXiv2022

MedSegDiff: Medical Image Segmentation with Diffusion Probabilistic Model

Diffusion probabilistic model (DPM) recently becomes one of the hottest topic in computer vision. Its image generation application such as Imagen, Latent Diffusion Models and Stable Diffusion have shown impressive generation capabilities, which aroused extensive discussion in the community. Many recent studies also found it is useful in many other vision tasks, like image deblurring, super-resolution and anomaly detection. Inspired by the success of DPM, we propose the first DPM based model toward general medical image segmentation tasks, which we named MedSegDiff. In order to enhance the step-wise regional attention in DPM for the medical image segmentation, we propose dynamic conditional encoding, which establishes the state-adaptive conditions for each sampling step. We further propose Feature Frequency Parser (FF-Parser), to eliminate the negative effect of high-frequency noise component in this process. We verify MedSegDiff on three medical segmentation tasks with different image modalities, which are optic cup segmentation over fundus images, brain tumor segmentation over MRI images and thyroid nodule segmentation over ultrasound images. The experimental results show that MedSegDiff outperforms state-of-the-art (SOTA) methods with considerable performance gap, indicating the generalization and effectiveness of the proposed model. Our code is released at https://github.com/WuJunde/MedSegDiff.

cs.CV↗

arXiv2022

SeATrans: Learning Segmentation-Assisted diagnosis model via Transformer

Clinically, the accurate annotation of lesions/tissues can significantly facilitate the disease diagnosis. For example, the segmentation of optic disc/cup (OD/OC) on fundus image would facilitate the glaucoma diagnosis, the segmentation of skin lesions on dermoscopic images is helpful to the melanoma diagnosis, etc. With the advancement of deep learning techniques, a wide range of methods proved the lesions/tissues segmentation can also facilitate the automated disease diagnosis models. However, existing methods are limited in the sense that they can only capture static regional correlations in the images. Inspired by the global and dynamic nature of Vision Transformer, in this paper, we propose Segmentation-Assisted diagnosis Transformer (SeATrans) to transfer the segmentation knowledge to the disease diagnosis network. Specifically, we first propose an asymmetric multi-scale interaction strategy to correlate each single low-level diagnosis feature with multi-scale segmentation features. Then, an effective strategy called SeA-block is adopted to vitalize diagnosis feature via correlated segmentation features. To model the segmentation-diagnosis interaction, SeA-block first embeds the diagnosis feature based on the segmentation information via the encoder, and then transfers the embedding back to the diagnosis feature space by a decoder. Experimental results demonstrate that SeATrans surpasses a wide range of state-of-the-art (SOTA) segmentation-assisted diagnosis methods on several disease diagnosis tasks.

cs.CV↗