Xuecheng Wu

Minghao Han, Yichen Liu, Yizhou Liu et al.

In computational pathology, understanding and generation have evolved along disparate paths: advanced understanding models already exhibit diagnostic-level competence, whereas generative models largely simulate pixels. Progress remains hindered by three coupled factors: the scarcity of large, high-quality image-text corpora; the lack of precise, fine-grained semantic control, which forces reliance on non-semantic cues; and terminological heterogeneity, where diverse phrasings for the same diagnostic concept impede reliable text conditioning. We introduce UniPath, a semantics-driven pathology image generation framework that leverages mature diagnostic understanding to enable controllable generation. UniPath implements Multi-Stream Control: a Raw-Text stream; a High-Level Semantics stream that uses learnable queries to a frozen pathology MLLM to distill paraphrase-robust Diagnostic Semantic Tokens and to expand prompts into diagnosis-aware attribute bundles; and a Prototype stream that affords component-level morphological control via a prototype bank. On the data front, we curate a 2.65M image-text corpus and a finely annotated, high-quality 68K subset to alleviate data scarcity. For a comprehensive assessment, we establish a four-tier evaluation hierarchy tailored to pathology. Extensive experiments demonstrate UniPath's SOTA performance, including a Patho-FID of 80.9 (51% better than the second-best) and fine-grained semantic control achieving 98.7% of the real-image. The dataset and code can be obtained from https://github.com/Hanminghao/UniPath.

Xuecheng Wu, Mengmeng Tian, Lanhang Zhai

Recently, the domestic COVID-19 epidemic situation is serious, but in public places, some people do not wear masks or wear masks incorrectly, which requires the relevant staff to instantly remind and supervise them to wear masks correctly. However, in the face of such an important and complicated work, it is very necessary to carry out automated mask-wearing detection in public places. This paper proposes a new mask-wearing detection method based on improved YOLOv4. Specifically, firstly, we add the Coordinate Attention Module to the backbone to coordinate feature fusion and representation. Secondly, we conduct a series of network structural improvements to enhance the model performance and robustness. Thirdly, we adaptively deploy the K-means clustering algorithm to make the nine anchor boxes more suitable for our NPMD dataset. The experiments show that the improved YOLOv4 performs better, exceeding the baseline by 4.06\% AP with a comparable speed of 64.37 FPS.

Wenxuan Jiang, Yuxin Zuo, Zijian Zhang et al.

In-Context Reinforcement Learning (ICRL) enables Large Language Models (LLMs) to learn online from external rewards directly within the context window. However, a central challenge in ICRL is reward estimation, as models typically lack access to ground-truths during inference. To address this limitation, we propose Test-Time Rethinking for In-Context Reinforcement Learning (TR-ICRL), a novel ICRL framework designed for both reasoning and knowledge-intensive tasks. TR-ICRL operates by first retrieving the most relevant instances from an unlabeled evaluation set for a given query. During each ICRL iteration, LLM generates a set of candidate answers for every retrieved instance. Next, a pseudo-label is derived from this set through majority voting. This label then serves as a proxy to give reward messages and generate formative feedbacks, guiding LLM through iterative refinement. In the end, this synthesized contextual information is integrated with the original query to form a comprehensive prompt, with the answer determining through a final round of majority voting. TR-ICRL is evaluated on mainstream reasoning and knowledge-intensive tasks, where it demonstrates significant performance gains. Remarkably, TR-ICRL improves Qwen2.5-7B by 21.23% on average on MedQA and even 137.59% on AIME2024. Extensive ablation studies and analyses further validate the effectiveness and robustness of our approach. Our code is available at https://github.com/pangpang-xuan/TR_ICRL.

Wei Xue, Mingcheng Li, Xuecheng Wu et al.

Vision-and-Language Navigation (VLN) requires agents to accurately perceive complex visual environments and reason over navigation instructions and histories. However, existing methods passively process redundant visual inputs and treat all historical contexts indiscriminately, resulting in inefficient perception and unfocused reasoning. To address these challenges, we propose \textbf{ProFocus}, a training-free progressive framework that unifies \underline{Pro}active Perception and \underline{Focus}ed Reasoning through collaboration between large language models (LLMs) and vision-language models (VLMs). For proactive perception, ProFocus transforms panoramic observations into structured ego-centric semantic maps, enabling the orchestration agent to identify missing visual information needed for reliable decision-making, and to generate targeted visual queries with corresponding focus regions that guide the perception agent to acquire the required observations. For focused reasoning, we propose Branch-Diverse Monte Carlo Tree Search (BD-MCTS) to identify top-$k$ high-value waypoints from extensive historical candidates. The decision agent focuses reasoning on the historical contexts associated with these waypoints, rather than considering all historical waypoints equally. Extensive experiments validate the effectiveness of ProFocus, achieving state-of-the-art performance among zero-shot methods on R2R and REVERIE benchmarks.

Dingkang Yang, Mingcheng Li, Xuecheng Wu et al.

Multimodal Sentiment Analysis (MSA) aims to predict sentiment from language, acoustic, and visual data in videos. However, imbalanced unimodal performance often leads to suboptimal fused representations. Existing approaches typically adopt fixed primary modality strategies to maximize dominant modality advantages, yet fail to adapt to dynamic variations in modality importance across different samples. Moreover, non-language modalities suffer from sequential redundancy and noise, degrading model performance when they serve as primary inputs. To address these issues, this paper proposes a modality optimization and dynamic primary modality selection framework (MODS). First, a Graph-based Dynamic Sequence Compressor (GDC) is constructed, which employs capsule networks and graph convolution to reduce sequential redundancy in acoustic/visual modalities. Then, we develop a sample-adaptive Primary Modality Selector (MSelector) for dynamic dominance determination. Finally, a Primary-modality-Centric Cross-Attention (PCCA) module is designed to enhance dominant modalities while facilitating cross-modal interaction. Extensive experiments on four benchmark datasets demonstrate that MODS outperforms state-of-the-art methods, achieving superior performance by effectively balancing modality contributions and eliminating redundant noise.

Hanshen Zhu, Yuliang Liu, Xuecheng Wu et al.

Visual Text Rendering (VTR) remains a critical challenge in text-to-image generation, where even advanced models frequently produce text with structural anomalies such as distortion, blurriness, and misalignment. However, we find that leading MLLMs and specialist OCR models largely fail to perceive these structural anomalies, creating a critical bottleneck for both VTR evaluation and RL-based optimization. As a result, even state-of-the-art generators (e.g., Seedream4.0, Qwen-Image) still struggle to render structurally faithful text. To address this, we propose TextPecker, a plug-and-play structural anomaly perceptive RL strategy that mitigates noisy reward signals and works with any textto-image generator. To enable this capability, we construct a recognition dataset with character-level structural-anomaly annotations and develop a stroke-editing synthesis engine to expand structural-error coverage. Experiments show that TextPecker consistently improves diverse text-to-image models; even on the well-optimized Qwen-Image, it significantly yields average gains of 4% in structural fidelity and 8.7% in semantic alignment for Chinese text rendering, establishing a new state-of-the-art in high-fidelity VTR. Our work fills a gap in VTR optimization, providing a foundational step towards reliable and structural faithful visual text generation.

Xuecheng Wu, Mengmeng Tian, Lanhang Zhai

With the fast development of artificial intelligence and short videos, emotion recognition in short videos has become one of the most important research topics in human-computer interaction. At present, most emotion recognition methods still stay in a single modality. However, in daily life, human beings will usually disguise their real emotions, which leads to the problem that the accuracy of single modal emotion recognition is relatively terrible. Moreover, it is not easy to distinguish similar emotions. Therefore, we propose a new approach denoted as ICANet to achieve multimodal short video emotion recognition by employing three different modalities of audio, video and optical flow, making up for the lack of a single modality and then improving the accuracy of emotion recognition in short videos. ICANet has a better accuracy of 80.77% on the IEMOCAP benchmark, exceeding the SOTA methods by 15.89%.

Junxiao Xue, Jie Wang, Xuecheng Wu et al.

Video content is rich in semantics and has the ability to evoke various emotions in viewers. In recent years, with the rapid development of affective computing and the explosive growth of visual data, affective video content analysis (AVCA) as an essential branch of affective computing has become a widely researched topic. In this study, we comprehensively review the development of AVCA over the past decade, particularly focusing on the most advanced methods adopted to address the three major challenges of video feature extraction, expression subjectivity, and multimodal feature fusion. We first introduce the widely used emotion representation models in AVCA and describe commonly used datasets. We summarize and compare representative methods in the following aspects: (1) unimodal AVCA models, including facial expression recognition and posture emotion recognition; (2) multimodal AVCA models, including feature fusion, decision fusion, and attention-based multimodal models; (3) model performance evaluation standards. Finally, we discuss future challenges and promising research directions, such as emotion recognition and public opinion analysis, human-computer interaction, and emotional intelligence.

Hao Feng, Wei Shi, Ke Zhang et al.

Document parsing has garnered widespread attention as vision-language models (VLMs) advance OCR capabilities. However, the field remains fragmented across dozens of specialized models with varying strengths, forcing users to navigate complex model selection and limiting system scalability. Moreover, existing two-stage approaches depend on axis-aligned bounding boxes for layout detection, failing to handle distorted or photographed documents effectively. To this end, we present Dolphin-v2, a two-stage document image parsing model that substantially improves upon the original Dolphin. In the first stage, Dolphin-v2 jointly performs document type classification (digital-born versus photographed) alongside layout analysis. For digital-born documents, it conducts finer-grained element detection with reading order prediction. In the second stage, we employ a hybrid parsing strategy: photographed documents are parsed holistically as complete pages to handle geometric distortions, while digital-born documents undergo element-wise parallel parsing guided by the detected layout anchors, enabling efficient content extraction. Compared with the original Dolphin, Dolphin-v2 introduces several crucial enhancements: (1) robust parsing of photographed documents via holistic page-level understanding, (2) finer-grained element detection (21 categories) with semantic attribute extraction such as author information and document metadata, and (3) code block recognition with indentation preservation, which existing systems typically lack. Comprehensive evaluations are conducted on DocPTBench, OmniDocBench, and our self-constructed RealDoc-160 benchmark. The results demonstrate substantial improvements: +14.78 points overall on the challenging OmniDocBench and 91% error reduction on photographed documents, while maintaining efficient inference through parallel processing.

Shi Chen, Xuecheng Wu, Heli Sun et al.

Affective Image Manipulation (AIM) aims to evoke specific emotions through targeted editing. Current image editing benchmarks primarily focus on object-level modifications in general scenarios, lacking the fine-grained granularity to capture affective dimensions. To bridge this gap, we introduce the first benchmark designed for AIM termed AIM-Bench. This benchmark is built upon a dual-path affective modeling scheme that integrates the Mikels emotion taxonomy with the Valence-Arousal-Dominance framework, enabling high-level semantic and fine-grained continuous manipulation. Through a hierarchical human-in-the-loop workflow, we finally curate 800 high-quality samples covering 8 emotional categories and 5 editing types. To effectively assess performance, we also design a composite evaluation suite combining rule-based and model-based metrics to holistically assess instruction consistency, aesthetics, and emotional expressiveness. Extensive evaluations reveal that current editing models face significant challenges, most notably a prevalent positivity bias, which stemming from inherent imbalances in training data distribution. To tackle this, we propose a scalable data engine utilizing an inverse repainting strategy to construct AIM-40k, a balanced instruction-tuning dataset comprising 40k samples. Concretely, we enhance raw affective images via generative redrawing to establish high-fidelity ground truths, and synthesize input images with divergent emotions and paired precise instructions. Fine-tuning a baseline model on AIM-40k yields a 9.15% relative improvement in overall performance, demonstrating the effectiveness of our AIM-40k. Our data and related code will be made open soon.

Xuecheng Wu, Heli Sun, Junxiao Xue et al.

Nowadays, short-form videos (SVs) are essential to web information acquisition and sharing in our daily life. The prevailing use of SVs to spread emotions leads to the necessity of conducting video emotion analysis (VEA) towards SVs. Considering the lack of SVs emotion data, we introduce a large-scale dataset named eMotions, comprising 27,996 videos. Meanwhile, we alleviate the impact of subjectivities on labeling quality by emphasizing better personnel allocations and multi-stage annotations. In addition, we provide the category-balanced and test-oriented variants through targeted data sampling. Some commonly used videos, such as facial expressions, have been well studied. However, it is still challenging to analysis the emotions in SVs. Since the broader content diversity brings more distinct semantic gaps and difficulties in learning emotion-related features, and there exists local biases and collective information gaps caused by the emotion inconsistence under the prevalently audio-visual co-expressions. To tackle these challenges, we present an end-to-end audio-visual baseline AV-CANet which employs the video transformer to better learn semantically relevant representations. We further design the Local-Global Fusion Module to progressively capture the correlations of audio-visual features. The EP-CE Loss is then introduced to guide model optimization. Extensive experimental results on seven datasets demonstrate the effectiveness of AV-CANet, while providing broad insights for future works. Besides, we investigate the key components of AV-CANet by ablation studies. Datasets and code will be fully open soon.

Junxiao Xue, Pavel Smirnov, Ziao Li et al.

We propose a real-time 3D human pose estimation and motion analysis method termed RePose for rehabilitation training. It is capable of real-time monitoring and evaluation of patients'motion during rehabilitation, providing immediate feedback and guidance to assist patients in executing rehabilitation exercises correctly. Firstly, we introduce a unified pipeline for end-to-end real-time human pose estimation and motion analysis using RGB video input from multiple cameras which can be applied to the field of rehabilitation training. The pipeline can help to monitor and correct patients'actions, thus aiding them in regaining muscle strength and motor functions. Secondly, we propose a fast tracking method for medical rehabilitation scenarios with multiple-person interference, which requires less than 1ms for tracking for a single frame. Additionally, we modify SmoothNet for real-time posture estimation, effectively reducing pose estimation errors and restoring the patient's true motion state, making it visually smoother. Finally, we use Unity platform for real-time monitoring and evaluation of patients' motion during rehabilitation, and to display the muscle stress conditions to assist patients with their rehabilitation training.

Xuecheng Wu, Junxiao Xue, Xinyi Yin et al.

Affective video facial analysis (AVFA) has emerged as a key research field for building emotion-aware intelligent systems, yet this field continues to suffer from limited data availability. In recent years, the self-supervised learning (SSL) technique of Masked Autoencoders (MAE) has gained momentum, with growing adaptations in its audio-visual contexts. While scaling has proven essential for breakthroughs in general multi-modal learning domains, its specific impact on AVFA remains largely unexplored. Another core challenge in this field is capturing both intra- and inter-modal correlations through scalable audio-visual representations. To tackle these issues, we propose AVF-MAE++, a family of audio-visual MAE models designed to efficiently investigate the scaling properties in AVFA while enhancing cross-modal correlation modeling. Our framework introduces a novel dual masking strategy across audio and visual modalities and strengthens modality encoders with a more holistic design to better support scalable pre-training. Additionally, we present the Iterative Audio-Visual Correlation Learning Module, which improves correlation learning within the SSL paradigm, bridging the limitations of previous methods. To support smooth adaptation and reduce overfitting risks, we further introduce a progressive semantic injection strategy, organizing the model training into three structured stages. Extensive experiments conducted on 17 datasets, covering three major AVFA tasks, demonstrate that AVF-MAE++ achieves consistent state-of-the-art performance across multiple benchmarks. Comprehensive ablation studies further highlight the importance of each proposed component and provide deeper insights into the design choices driving these improvements. Our code and models have been publicly released at Github.

Siyu Yan, Long Zeng, Xuecheng Wu et al.

As large language models~(LLMs) become widely adopted, ensuring their alignment with human values is crucial to prevent jailbreaks where adversaries manipulate models to produce harmful content. While most defenses target single-turn attacks, real-world usage often involves multi-turn dialogues, exposing models to attacks that exploit conversational context to bypass safety measures. We introduce MUSE, a comprehensive framework tackling multi-turn jailbreaks from both attack and defense angles. For attacks, we propose MUSE-A, a method that uses frame semantics and heuristic tree search to explore diverse semantic trajectories. For defense, we present MUSE-D, a fine-grained safety alignment approach that intervenes early in dialogues to reduce vulnerabilities. Extensive experiments on various models show that MUSE effectively identifies and mitigates multi-turn vulnerabilities. Code is available at \href{https://github.com/yansiyu02/MUSE}{https://github.com/yansiyu02/MUSE}.

Fei Wang, Xuecheng Wu, Zheng Zhang et al.

The powerful generative capabilities of diffusion models have significantly advanced the field of image synthesis, enhancing both full image generation and inpainting-based image editing. Despite their remarkable advancements, diffusion models also raise concerns about potential misuse for malicious purposes. However, existing approaches struggle to identify images generated by diffusion-based inpainting models, even when similar inpainted images are included in their training data. To address this challenge, we propose a novel detection method based on End-to-end denoising diffusion (End4). Specifically, End4 designs a denoising reconstruction model to improve the alignment degree between the latent spaces of the reconstruction and detection processes, thus reconstructing features that are more conducive to detection. Meanwhile, it leverages a Scale-aware Pyramid-like Fusion Module (SPFM) that refines local image features under the guidance of attention pyramid layers at different scales, enhancing feature discriminability. Additionally, to evaluate detection performance on inpainted images, we establish a comprehensive benchmark comprising images generated from five distinct masked regions. Extensive experiments demonstrate that our End4 effectively generalizes to unseen masking patterns and remains robust under various perturbations. Our code and dataset will be released soon.

Junxiao Xue, Xiaozhen Liu, Xuecheng Wu et al.

Estimating spoken content from silent videos is crucial for applications in Assistive Technology (AT) and Augmented Reality (AR). However, accurately mapping lip movement sequences in videos to words poses significant challenges due to variability across sequences and the uneven distribution of information within each sequence. To tackle this, we introduce InfoSyncNet, a non-uniform sequence modeling network enhanced by tailored data augmentation techniques. Central to InfoSyncNet is a non-uniform quantization module positioned between the encoder and decoder, enabling dynamic adjustment to the network's focus and effectively handling the natural inconsistencies in visual speech data. Additionally, multiple training strategies are incorporated to enhance the model's capability to handle variations in lighting and the speaker's orientation. Comprehensive experiments on the LRW and LRW1000 datasets confirm the superiority of InfoSyncNet, achieving new state-of-the-art accuracies of 92.0% and 60.7% Top-1 ACC. The code is available for download (see comments).

Junxiao Xue, Quan Deng, Tingqi Hu et al.

Despite recent progress in multimodal large language models (MLLMs), reliable visual question answering in aerial scenes remains challenging. In such scenes, task-critical evidence is often carried by small objects, explicit quantities, coarse locations, and inter-object relations, whereas conventional dense visual-token representations are not well aligned with these structured semantics. To address this interface mismatch, we propose AeroRAG, a scene-graph-guided multimodal retrieval-augmented generation framework for visual question answering. The framework first converts an input image into structured visual knowledge, including object categories, quantities, spatial locations, and semantic relations, and then retrieves query-relevant semantic chunks to construct compact prompts for a text-based large language model. Rather than relying on direct reasoning over dense visual tokens, our method introduces a more explicit intermediate interface between perception and language reasoning. Experiments on the AUG aerial dataset and the general-domain VG-150 benchmark show consistent improvements over six strong MLLM baselines, with the largest gains observed in dense aerial scenes and relation-sensitive reasoning. We further evaluate the framework on VQAv2 to verify that the proposed interface remains compatible with standard visual reasoning settings. These results suggest that structured retrieval is a practical design direction for deployment-oriented and grounded visual reasoning systems.

Chi Ding, Junxiao Xue, Xinyi Yin et al.

Long-Tailed distributions are pervasive in remote sensing due to the inherently imbalanced occurrence of grounded objects. However, a critical challenge remains largely overlooked, i.e., disentangling hard tail data samples from noisy ambiguous ones. Conventional methods often indiscriminately emphasize all low-confidence samples, leading to overfitting on noisy data. To bridge this gap, building upon Evidential Deep Learning, we propose a model-agnostic uncertainty-aware framework termed DUAL, which dynamically disentangles prediction uncertainty into Epistemic Uncertainty (EU) and Aleatoric Uncertainty (AU). Specifically, we introduce EU as an indicator of sample scarcity to guide a reweighting strategy for hard-to-learn tail samples, while leveraging AU to quantify data ambiguity, employing an adaptive label smoothing mechanism to suppress the impact of noise. Extensive experiments on multiple datasets across various backbones demonstrate the effectiveness and generalization of our framework, surpassing strong baselines such as TGN and SADE. Ablation studies provide further insights into the crucial choices of our design.

Xinyi Yin, Yiduo Wang, Tingqi Hu et al.

Affective image editing (AIE) aims to edit visual content to evoke target emotions. However, existing methods often overlook inference efficiency and predominantly depend on discrete emotion representations, which to some extent limits their practical applicability and makes it challenging to capture complex and subtle human emotions. To tackle these gaps, we propose MooD, the first framework that directly leverages continuous Valence-Arousal (VA) values for fine-grained and efficient AIE. Specifically, we first introduce a VA-Aware retrieval strategy to bridge vague affective values and concrete visual semantics. Building upon this, MooD integrates visual transfer and semantic guidance to achieve controllable AIE. Furthermore, we construct AffectSet, a VA-annotated dataset to support model optimization and evaluation. Extensive qualitative and quantitative experimental results demonstrate that our MooD achieves superior performance in both affective controllability and visual fidelity while maintaining high efficiency. A series of ablation studies further reveal the crucial factors of our design. Our code and data will be made publicly open soon.

Shuhao Han, Haotian Fan, Fangyuan Kong et al.

This paper reports on the NTIRE 2025 challenge on Text to Image (T2I) generation model quality assessment, which will be held in conjunction with the New Trends in Image Restoration and Enhancement Workshop (NTIRE) at CVPR 2025. The aim of this challenge is to address the fine-grained quality assessment of text-to-image generation models. This challenge evaluates text-to-image models from two aspects: image-text alignment and image structural distortion detection, and is divided into the alignment track and the structural track. The alignment track uses the EvalMuse-40K, which contains around 40K AI-Generated Images (AIGIs) generated by 20 popular generative models. The alignment track has a total of 371 registered participants. A total of 1,883 submissions are received in the development phase, and 507 submissions are received in the test phase. Finally, 12 participating teams submitted their models and fact sheets. The structure track uses the EvalMuse-Structure, which contains 10,000 AI-Generated Images (AIGIs) with corresponding structural distortion mask. A total of 211 participants have registered in the structure track. A total of 1155 submissions are received in the development phase, and 487 submissions are received in the test phase. Finally, 8 participating teams submitted their models and fact sheets. Almost all methods have achieved better results than baseline methods, and the winning methods in both tracks have demonstrated superior prediction performance on T2I model quality assessment.

Kedi Chen, Dezhao Ruan, Yuhao Dan et al.

Reasoning is an important task for large language models (LLMs). Among all the reasoning paradigms, inductive reasoning is one of the fundamental types, which is characterized by its particular-to-general thinking process and the non-uniqueness of its answers. The inductive mode is crucial for knowledge generalization and aligns better with human cognition, so it is a fundamental mode of learning, hence attracting increasing interest. Despite the importance of inductive reasoning, there is no systematic summary of it. Therefore, this paper presents the first comprehensive survey of inductive reasoning for LLMs. First, methods for improving inductive reasoning are categorized into three main areas: post-training, test-time scaling, and data augmentation. Then, current benchmarks of inductive reasoning are summarized, and a unified sandbox-based evaluation approach with the observation coverage metric is derived. Finally, we offer some analyses regarding the source of inductive ability and how simple model architectures and data help with inductive tasks, providing a solid foundation for future research.

Zijian Zhang, Xuhui Zheng, Xuecheng Wu et al.

While text-to-image (T2I) generation models have achieved remarkable progress in recent years, existing evaluation methodologies for vision-language alignment still struggle with the fine-grained semantic matching. Current approaches based on global similarity metrics often overlook critical token-level correspondences between textual descriptions and visual content. To this end, we present TokenFocus-VQA, a novel evaluation framework that leverages Large Vision-Language Models (LVLMs) through visual question answering (VQA) paradigm with position-specific probability optimization. Our key innovation lies in designing a token-aware loss function that selectively focuses on probability distributions at pre-defined vocabulary positions corresponding to crucial semantic elements, enabling precise measurement of fine-grained semantical alignment. The proposed framework further integrates ensemble learning techniques to aggregate multi-perspective assessments from diverse LVLMs architectures, thereby achieving further performance enhancement. Evaluated on the NTIRE 2025 T2I Quality Assessment Challenge Track 1, our TokenFocus-VQA ranks 2nd place (0.8445, only 0.0001 lower than the 1st method) on public evaluation and 2nd place (0.8426) on the official private test set, demonstrating superiority in capturing nuanced text-image correspondences compared to conventional evaluation methods.

Zijian Zhang, Xuecheng Wu, Danlei Huang et al.

Driven by the rapid progress in vision-language models (VLMs), the responsible behavior of large-scale multimodal models has become a prominent research area, particularly focusing on hallucination detection and factuality checking. In this paper, we present the solution for the two tracks of Responsible AI challenge. Inspirations from the general domain demonstrate that a smaller distilled VLM can often outperform a larger VLM that is directly tuned on downstream tasks, while achieving higher efficiency. We thus jointly tackle two tasks from the perspective of knowledge distillation and propose a progressive hybrid knowledge distillation framework termed HKD4VLM. Specifically, the overall framework can be decomposed into Pyramid-like Progressive Online Distillation and Ternary-Coupled Refinement Distillation, hierarchically moving from coarse-grained knowledge alignment to fine-grained refinement. Besides, we further introduce the mapping shift-enhanced inference and diverse augmentation strategies to enhance model performance and robustness. Extensive experimental results demonstrate the effectiveness of our HKD4VLM. Ablation studies provide insights into the critical design choices driving performance gains.

Xuecheng Wu, Jiaxing Liu, Danlei Huang et al.

Visual-Interleaved Chain-of-Thought (VI-CoT) enables MLLMs to continually update their understanding and decisions based on step-wise intermediate visual states (IVS), much like a human would, which demonstrates impressive success in various tasks, thereby leading to emerged advancements in related benchmarks. Despite promising progress, current benchmarks provide models with relatively fixed IVS, rather than free-style IVS, whch might forcibly distort the original thinking trajectories, failing to evaluate their intrinsic reasoning capabilities. More importantly, existing benchmarks neglect to systematically explore the impact factors that IVS would impart to untamed reasoning performance. To tackle above gaps, we introduce a specialized benchmark termed ViC-Bench, consisting of four representive tasks: maze navigation, jigsaw puzzle, embodied long-horizon planning, and complex counting, where each task has dedicated free-style IVS generation pipeline supporting function calls. To systematically examine VI-CoT capability, we propose a thorough evaluation suite incorporating a progressive three-stage strategy with targeted new metrics. Besides, we establish Incremental Prompting Information Injection (IPII) strategy to ablatively explore the prompting factors for VI-CoT. We extensively conduct evaluations for 18 advanced MLLMs, revealing key insights into their VI-CoT capability. Our proposed benchmark is publicly open at Huggingface.

Junbo Wang, Liangyu Fu, Yuke Li et al.

Micro-expression recognition can obtain the real emotion of the individual at the current moment. Although deep learning-based methods, especially Transformer-based methods, have achieved impressive results, these methods have high computational complexity due to the large number of tokens in the multi-head self-attention. In addition, the existing micro-expression datasets are small-scale, which makes it difficult for Transformer-based models to learn effective micro-expression representations. Therefore, we propose a novel Efficient Patch tokenization, Integration and Representation framework (EPIR), which can balance high recognition performance and low computational complexity. Specifically, we first propose a dual norm shifted tokenization (DNSPT) module to learn the spatial relationship between neighboring pixels in the face region, which is implemented by a refined spatial transformation and dual norm projection. Then, we propose a token integration module to integrate partial tokens among multiple cascaded Transformer blocks, thereby reducing the number of tokens without information loss. Furthermore, we design a discriminative token extractor, which first improves the attention in the Transformer block to reduce the unnecessary focus of the attention calculation on self-tokens, and uses the dynamic token selection module (DTSM) to select key tokens, thereby capturing more discriminative micro-expression representations. We conduct extensive experiments on four popular public datasets (i.e., CASME II, SAMM, SMIC, and CAS(ME)3. The experimental results show that our method achieves significant performance gains over the state-of-the-art methods, such as 9.6% improvement on the CAS(ME)$^3$ dataset in terms of UF1 and 4.58% improvement on the SMIC dataset in terms of UAR metric.

Junbo Wang, Liangyu Fu, Yuke Li et al.

Current video captioning methods usually use an encoder-decoder structure to generate text autoregressively. However, autoregressive methods have inherent limitations such as slow generation speed and large cumulative error. Furthermore, the few non-autoregressive counterparts suffer from deficiencies in generation quality due to the lack of sufficient multimodal interaction modeling. Therefore, we propose a non-autoregressive framework based on Diffusion model for Video Captioning (DiffVC) to address these issues. Its parallel decoding can effectively solve the problems of generation speed and cumulative error. At the same time, our proposed discriminative conditional Diffusion Model can generate higher-quality textual descriptions. Specifically, we first encode the video into a visual representation. During training, Gaussian noise is added to the textual representation of the ground-truth caption. Then, a new textual representation is generated via the discriminative denoiser with the visual representation as a conditional constraint. Finally, we input the new textual representation into a non-autoregressive language model to generate captions. During inference, we directly sample noise from the Gaussian distribution for generation. Experiments on MSVD, MSR-VTT, and VATEX show that our method can outperform previous non-autoregressive methods and achieve comparable performance to autoregressive methods, e.g., it achieved a maximum improvement of 9.9 on the CIDEr and improvement of 2.6 on the B@4, while having faster generation speed. The source code will be available soon.

Xuecheng Wu, Dingkang Yang, Danlei Huang et al.

Short-form videos (SVs) have become a vital part of our online routine for acquiring and sharing information. Their multimodal complexity poses new challenges for video analysis, highlighting the need for video emotion analysis (VEA) within the community. Given the limited availability of SVs emotion data, we introduce eMotions, a large-scale dataset consisting of 27,996 videos with full-scale annotations. To ensure quality and reduce subjective bias, we emphasize better personnel allocation and propose a multi-stage annotation procedure. Additionally, we provide the category-balanced and test-oriented variants through targeted sampling to meet diverse needs. While there have been significant studies on videos with clear emotional cues (e.g., facial expressions), analyzing emotions in SVs remains a challenging task. The challenge arises from the broader content diversity, which introduces more distinct semantic gaps and complicates the representations learning of emotion-related features. Furthermore, the prevalence of audio-visual co-expressions in SVs leads to the local biases and collective information gaps caused by the inconsistencies in emotional expressions. To tackle this, we propose AV-CANet, an end-to-end audio-visual fusion network that leverages video transformer to capture semantically relevant representations. We further introduce the Local-Global Fusion Module designed to progressively capture the correlations of audio-visual features. Besides, EP-CE Loss is constructed to globally steer optimizations with tripolar penalties. Extensive experiments across three eMotions-related datasets and four public VEA datasets demonstrate the effectiveness of our proposed AV-CANet, while providing broad insights for future research. Moreover, we conduct ablation studies to examine the critical components of our method. Dataset and code will be made available at Github.

Fengjian Xue, Xuecheng Wu, Heli Sun et al.

Facial expression image editing requires fine-grained control to strictly preserve human identity and background while precisely manipulating expression. However, existing editing benchmarks primarily focus on general scenarios, lacking high-quality facial images and corresponding editing instructions. Furthermore, current evaluation metrics exhibit systemic biases in this task, often favoring lazy editing or overfit editing. To bridge these gaps, we propose FED-Bench, a comprehensive benchmark featuring rigorous testing and an accurate evaluation suite. First, we carefully construct a benchmark of 747 triplets through a cascaded and scalable pipeline, each comprising an original image, an editing instruction, and a ground-truth image for precise evaluation. Second, we introduce FED-Score, a cross-granularity evaluation protocol that disentangles assessment into three dimensions: Alignment for verifying instruction following, Fidelity for testing image quality and identity preservation, and Relative Expression Gain for quantifying the magnitude of expression changes, effectively mitigating the aforementioned evaluation biases. Third, we benchmark 18 image editing models, revealing that current approaches struggle to simultaneously achieve high fidelity and accurate expression manipulation, with fine-grained instruction following identified as the primary bottleneck. Finally, leveraging the scalable characteristic of introduced benchmark engine, we provide a 20k+ in-the-wild facial training set and demonstrate its effectiveness by fine-tuning a baseline model that achieves significant performance gains. Our benchmark and related code will be made publicly open soon.

Xuechun Liu, Heli Sun, Xuecheng Wu et al.

Multivariate time series anomaly detection (MTSAD) aims to accurately identify and localize complex abnormal patterns in the large-scale industrial control systems. While existing approaches excel in recognizing the distinct patterns under the low-dimensional scenarios, they often fail to robustly capture long-range spatiotemporal dependencies when learning representations from the high-dimensional noisy time series. To address these limitations, we propose DARTs, a robust long short-term dual-path framework with window-aware spatiotemporal soft fusion mechanism, which can be primarily decomposed into three complementary components. Specifically, in the short-term path, we introduce a Multi-View Sparse Graph Learner and a Diffusion Multi-Relation Graph Unit that collaborate to adaptively capture hierarchical discriminative short-term spatiotemporal patterns in the high-noise time series. While in the long-term path, we design a Multi-Scale Spatiotemporal Graph Constructor to model salient long-term dynamics within the high-dimensional representation space. Finally, a window-aware spatiotemporal soft-fusion mechanism is introduced to filter the residual noise while seamlessly integrating anomalous patterns. Extensive qualitative and quantitative experimental results across mainstream datasets demonstrate the superiority and robustness of our proposed DARTs. A series of ablation studies are also conducted to explore the crucial design factors of our proposed components. Our code and model will be made publicly open soon.

Xiangyan Kong, Xuecheng Wu, Xiongwei Zhao et al.

V2X prediction can alleviate perception incompleteness caused by limited line of sight through fusing trajectory data from infrastructure and vehicles, which is crucial to traffic safety and efficiency. However, in dense traffic scenarios, frequent identity switching of targets hinders cross-view association and fusion. Meanwhile, multi-source information tends to generate redundant interactions during the encoding stage, and traditional vehicle-centric encoding leads to large amounts of repetitive historical trajectory feature encoding, degrading real-time inference performance. To address these challenges, we propose V2X-RECT, a trajectory prediction framework designed for high-density environments. It enhances data association consistency, reduces redundant interactions, and reuses historical information to enable more efficient and accurate prediction. Specifically, we design a multi-source identity matching and correction module that leverages multi-view spatiotemporal relationships to achieve stable and consistent target association, mitigating the adverse effects of mismatches on trajectory encoding and cross-view feature fusion. Then we introduce traffic signal-guided interaction module, encoding trend of traffic light changes as features and exploiting their role in constraining spatiotemporal passage rights to accurately filter key interacting vehicles, while capturing the dynamic impact of signal changes on interaction patterns. Furthermore, a local spatiotemporal coordinate encoding enables reusable features of historical trajectories and map, supporting parallel decoding and significantly improving inference efficiency. Extensive experimental results across V2X-Seq and V2X-Traj datasets demonstrate that our V2X-RECT achieves significant improvements compared to SOTA methods, while also enhancing robustness and inference efficiency across diverse traffic densities.

Kedi Chen, Zhikai Lei, Xu Guo et al.

Large language models (LLMs) make remarkable progress in reasoning tasks. Among different reasoning modes, inductive reasoning, due to its better alignment with human learning, attracts increasing interest. However, research on inductive reasoning faces certain challenges. First, existing inductive data mostly focuses on superficial regularities while lacking more complex internal patterns. Second, current works merely prompt LLMs or finetune on simple prompt-response pairs, but do not provide precise thinking processes nor implement difficulty control. Unlike previous work, we address these challenges by introducing \textit{CodeSeq}, a synthetic post-training dataset built from number sequences. We package number sequences into algorithmic problems to discover their general terms, defining a general term generation (GTG) task correspondingly. Our pipeline generates supervised finetuning data by reflecting on failed test cases and incorporating iterative corrections, thereby teaching LLMs to learn autonomous case generation and self-checking. Additionally, it leverages reinforcement learning with a novel Case-Synergy Solvability Scaling Reward based on both solvability, estimated from the problem pass rate, and the success rate of self-directed case generation, enabling models to learn more effectively from both successes and failures. Experimental results show that the models trained with \textit{CodeSeq} improve on various reasoning tasks and can preserve the models' OOD performance.

Junxiao Xue, Quan Deng, Xuecheng Wu et al.

Remote sensing change understanding (RSCU) is essential for analyzing remote sensing images and understanding how human activities affect the environment. However, existing datasets lack deep understanding and interactions in the diverse change captioning, counting, and localization tasks. To tackle these gaps, we construct ChangeIMTI, a new large-scale interactive multi-task instruction dataset that encompasses four complementary tasks including change captioning, binary change classification, change counting, and change localization. Building upon this new dataset, we further design a novel vision-guided vision-language model (ChangeVG) with dual-granularity awareness for bi-temporal remote sensing images (i.e., two remote sensing images of the same area at different times). The introduced vision-guided module is a dual-branch architecture that synergistically combines fine-grained spatial feature extraction with high-level semantic summarization. These enriched representations further serve as the auxiliary prompts to guide large vision-language models (VLMs) (e.g., Qwen2.5-VL-7B) during instruction tuning, thereby facilitating the hierarchical cross-modal learning. We extensively conduct experiments across four tasks to demonstrate the superiority of our approach. Remarkably, on the change captioning task, our method outperforms the strongest method Semantic-CC by 1.39 points on the comprehensive S*m metric, which integrates the semantic similarity and descriptive accuracy to provide an overall evaluation of change caption. Moreover, we also perform a series of ablation studies to examine the critical components of our method.

Haiyang Yu, Yuchuan Wu, Fan Shi et al.

Chinese ancient documents, invaluable carriers of millennia of Chinese history and culture, hold rich knowledge across diverse fields but face challenges in digitization and understanding, i.e., traditional methods only scan images, while current Vision-Language Models (VLMs) struggle with their visual and linguistic complexity. Existing document benchmarks focus on English printed texts or simplified Chinese, leaving a gap for evaluating VLMs on ancient Chinese documents. To address this, we present AncientDoc, the first benchmark for Chinese ancient documents, designed to assess VLMs from OCR to knowledge reasoning. AncientDoc includes five tasks (page-level OCR, vernacular translation, reasoning-based QA, knowledge-based QA, linguistic variant QA) and covers 14 document types, over 100 books, and about 3,000 pages. Based on AncientDoc, we evaluate mainstream VLMs using multiple metrics, supplemented by a human-aligned large language model for scoring.

Yifan Wang, Binbin Liu, Fengze Liu et al.

The data mixture used in the pre-training of a language model is a cornerstone of its final performance. However, a static mixing strategy is suboptimal, as the model's learning preferences for various data domains shift dynamically throughout training. Crucially, observing these evolving preferences in a computationally efficient manner remains a significant challenge. To address this, we propose TiKMiX, a method that dynamically adjusts the data mixture according to the model's evolving preferences. TiKMiX introduces Group Influence, an efficient metric for evaluating the impact of data domains on the model. This metric enables the formulation of the data mixing problem as a search for an optimal, influence-maximizing distribution. We solve this via two approaches: TiKMiX-D for direct optimization, and TiKMiX-M, which uses a regression model to predict a superior mixture. We trained models with different numbers of parameters, on up to 1 trillion tokens. TiKMiX-D exceeds the performance of state-of-the-art methods like REGMIX while using just 20% of the computational resources. TiKMiX-M leads to an average performance gain of 2% across 9 downstream benchmarks. Our experiments reveal that a model's data preferences evolve with training progress and scale, and we demonstrate that dynamically adjusting the data mixture based on Group Influence, a direct measure of these preferences, significantly improves performance by mitigating the underdigestion of data seen with static ratios.

Liangyu Fu, Xuecheng Wu, Danlei Huang et al.

Micro-expressions recognition (MER) has essential application value in many fields, but the short duration and low intensity of micro-expressions (MEs) bring considerable challenges to MER. The current MER methods in deep learning mainly include three data loading methods: static images, dynamic image sequence, and a combination of the two streams. How to effectively extract MEs' fine-grained and spatiotemporal features has been difficult to solve. This paper proposes a new MER method based on multi-task learning and hierarchical attention, which fully extracts MEs' omni-directional features by merging 2D and 3D CNNs. The fusion model consists of a 2D CNN AMNet2D and a 3D CNN AMNet3D, with similar structures consisting of a shared backbone network Resnet18 and attention modules. During training, the model adopts different data loading methods to adapt to two specific networks respectively, jointly trains on the tasks of MER and facial action unit detection (FAUD), and adopts the parameter hard sharing for information association, which further improves the effect of the MER task, and the final fused model is called FAMNet. Extensive experimental results show that our proposed FAMNet significantly improves task performance. On the SAMM, CASME II and MMEW datasets, FAMNet achieves 83.75% (UAR) and 84.03% (UF1). Furthermore, on the challenging CAS(ME)$^3$ dataset, FAMNet achieves 51% (UAR) and 43.42% (UF1).

Xinyi Yin, Wenbo Yuan, Xuecheng Wu et al.

Abnormal Human Behavior Detection (AHBD) under special scenarios is becoming increasingly crucial. While YOLO-based detection methods excel in real-time tasks, they remain hindered by challenges including small objects, task conflicts, and multi-scale fusion in AHBD. To tackle them, we propose TACR-YOLO, a new real-time framework for AHBD. We introduce a Coordinate Attention Module to enhance small object detection, a Task-Aware Attention Module to deal with classification-regression conflicts, and a Strengthen Neck Network for refined multi-scale fusion, respectively. In addition, we optimize Anchor Box sizes using K-means clustering and deploy DIoU-Loss to improve bounding box regression. The Personnel Anomalous Behavior Detection (PABD) dataset, which includes 8,529 samples across four behavior categories, is also presented. Extensive experimental results indicate that TACR-YOLO achieves 91.92% mAP on PABD, with competitive speed and robustness. Ablation studies highlight the contribution of each improvement. This work provides new insights for abnormal behavior detection under special scenarios, advancing its progress.

Junxiao Xue, Xiaozhen Liu, Jie Wang et al.

Existing emotion recognition methods mainly focus on enhancing performance by employing complex deep models, typically resulting in significantly higher model complexity. Although effective, it is also crucial to ensure the reliability of the final decision, especially for noisy, corrupted and out-of-distribution data. To this end, we propose a novel emotion recognition method called trusted emotion recognition (TER), which utilizes uncertainty estimation to calculate the confidence value of predictions. TER combines the results from multiple modalities based on their confidence values to output the trusted predictions. We also provide a new evaluation criterion to assess the reliability of predictions. Specifically, we incorporate trusted precision and trusted recall to determine the trusted threshold and formulate the trusted Acc. and trusted F1 score to evaluate the model's trusted performance. The proposed framework combines the confidence module that accordingly endows the model with reliability and robustness against possible noise or corruption. The extensive experimental results validate the effectiveness of our proposed model. The TER achieves state-of-the-art performance on the Music-video, achieving 82.40% Acc. In terms of trusted performance, TER outperforms other methods on the IEMOCAP and Music-video, achieving trusted F1 scores of 0.7511 and 0.9035, respectively.

Junxiao Xue, Xiaozhen Liu, Xuecheng Wu et al.

Audio-visual speech recognition (AVSR) combines audio-visual modalities to improve speech recognition, especially in noisy environments. However, most existing methods deploy the unidirectional enhancement or symmetric fusion manner, which limits their capability to capture heterogeneous and complementary correlations of audio-visual data-especially under asymmetric information conditions. To tackle these gaps, we introduce a new AVSR framework termed AD-AVSR based on bidirectional modality enhancement. Specifically, we first introduce the audio dual-stream encoding strategy to enrich audio representations from multiple perspectives and intentionally establish asymmetry to support subsequent cross-modal interactions. The enhancement process involves two key components, Audio-aware Visual Refinement Module for enhanced visual representations under audio guidance, and Cross-modal Noise Suppression Masking Module which refines audio representations using visual cues, collaboratively leading to the closed-loop and bidirectional information flow. To further enhance correlation robustness, we adopt a threshold-based selection mechanism to filter out irrelevant or weakly correlated audio-visual pairs. Extensive experimental results on the LRS2 and LRS3 datasets indicate that our AD-AVSR consistently surpasses SOTA methods in both performance and noise robustness, highlighting the effectiveness of our model design.

Xuecheng Wu, Danlei Huang, Heli Sun et al.

Advances in Generative AI have made video-level deepfake detection increasingly challenging, exposing the limitations of current detection techniques. In this paper, we present HOLA, our solution to the Video-Level Deepfake Detection track of 2025 1M-Deepfakes Detection Challenge. Inspired by the success of large-scale pre-training in the general domain, we first scale audio-visual self-supervised pre-training in the multimodal video-level deepfake detection, which leverages our self-built dataset of 1.81M samples, thereby leading to a unified two-stage framework. To be specific, HOLA features an iterative-aware cross-modal learning module for selective audio-visual interactions, hierarchical contextual modeling with gated aggregations under the local-global perspective, and a pyramid-like refiner for scale-aware cross-grained semantic enhancements. Moreover, we propose the pseudo supervised singal injection strategy to further boost model performance. Extensive experiments across expert models and MLLMs impressivly demonstrate the effectiveness of our proposed HOLA. We also conduct a series of ablation studies to explore the crucial design factors of our introduced components. Remarkably, our HOLA ranks 1st, outperforming the second by 0.0476 AUC on the TestA set.

Xuecheng Wu, Junxiao Xue, Liangyu Fu et al.

Recent research on real-time object detectors (e.g., YOLO series) has demonstrated the effectiveness of attention mechanisms for elevating model performance. Nevertheless, existing methods neglect to unifiedly deploy hierarchical attention mechanisms to construct a more discriminative YOLO head which is enriched with more useful intermediate features. To tackle this gap, this work aims to leverage multiple attention mechanisms to hierarchically enhance the triple discriminative awareness of the YOLO detection head and complementarily learn the coordinated intermediate representations, resulting in a new series detectors denoted 3A-YOLO. Specifically, we first propose a new head denoted TDA-YOLO Module, which unifiedly enhance the representations learning of scale-awareness, spatial-awareness, and task-awareness. Secondly, we steer the intermediate features to coordinately learn the inter-channel relationships and precise positional information. Finally, we perform neck network improvements followed by introducing various tricks to boost the adaptability of 3A-YOLO. Extensive experiments across COCO and VOC benchmarks indicate the effectiveness of our detectors.