Meiqi Wu

Zhe Huang, Hao Wen, Aiming Hao et al.

Multimodal Large Language Models (MLLMs) have made remarkable progress in video understanding. However, they suffer from a critical vulnerability: an over-reliance on language priors, which can lead to visual ungrounded hallucinations, especially when processing counterfactual videos that defy common sense. This limitation, stemming from the intrinsic data imbalance between text and video, is challenging to address due to the substantial cost of collecting and annotating counterfactual data. To address this, we introduce DualityForge, a novel counterfactual data synthesis framework that employs controllable, diffusion-based video editing to transform real-world videos into counterfactual scenarios. By embedding structured contextual information into the video editing and QA generation processes, the framework automatically produces high-quality QA pairs together with original-edited video pairs for contrastive training. Based on this, we build DualityVidQA, a large-scale video dataset designed to reduce MLLM hallucinations. In addition, to fully exploit the contrastive nature of our paired data, we propose Duality-Normalized Advantage Training (DNA-Train), a two-stage SFT-RL training regime where the RL phase applies pair-wise $\ell_1$ advantage normalization, thereby enabling a more stable and efficient policy optimization. Experiments on DualityVidQA-Test demonstrate that our method substantially reduces model hallucinations on counterfactual videos, yielding a relative improvement of 24.0% over the Qwen2.5-VL-7B baseline. Moreover, our approach achieves significant gains across both hallucination and general-purpose benchmarks, indicating strong generalization capability. We will open-source our dataset and code.

Xuchen Li, Shiyu Hu, Xiaokun Feng et al.

Visual Language Tracking (VLT) enhances tracking by mitigating the limitations of relying solely on the visual modality, utilizing high-level semantic information through language. This integration of the language enables more advanced human-machine interaction. The essence of interaction is cognitive alignment, which typically requires multiple information exchanges, especially in the sequential decision-making process of VLT. However, current VLT benchmarks do not account for multi-round interactions during tracking. They provide only an initial text and bounding box (bbox) in the first frame, with no further interaction as tracking progresses, deviating from the original motivation of the VLT task. To address these limitations, we propose a novel and robust benchmark, VLT-MI (Visual Language Tracking with Multi-modal Interaction), which introduces multi-round interaction into the VLT task for the first time. (1) We generate diverse, multi-granularity texts for multi-round, multi-modal interaction based on existing mainstream VLT benchmarks using DTLLM-VLT, leveraging the world knowledge of LLMs. (2) We propose a new VLT interaction paradigm that achieves multi-round interaction through text updates and object recovery. When multiple tracking failures occur, we provide the tracker with more aligned texts and corrected bboxes through interaction, thereby expanding the scope of VLT downstream tasks. (3) We conduct comparative experiments on both traditional VLT benchmarks and VLT-MI, evaluating and analyzing the accuracy and robustness of trackers under the interactive paradigm. This work offers new insights and paradigms for the VLT task, enabling a fine-grained evaluation of multi-modal trackers. We believe this approach can be extended to additional datasets in the future, supporting broader evaluations and comparisons of video-language model capabilities.

Chubin Chen, Sujie Hu, Jiashu Zhu et al.

Recent studies have demonstrated significant progress in aligning text-to-image diffusion models with human preference via Reinforcement Learning from Human Feedback. However, while existing methods achieve high scores on automated reward metrics, they often lead to Preference Mode Collapse (PMC)-a specific form of reward hacking where models converge on narrow, high-scoring outputs (e.g., images with monolithic styles or pervasive overexposure), severely degrading generative diversity. In this work, we introduce and quantify this phenomenon, proposing DivGenBench, a novel benchmark designed to measure the extent of PMC. We posit that this collapse is driven by over-optimization along the reward model's inherent biases. Building on this analysis, we propose Directional Decoupling Alignment (D$^2$-Align), a novel framework that mitigates PMC by directionally correcting the reward signal. Specifically, our method first learns a directional correction within the reward model's embedding space while keeping the model frozen. This correction is then applied to the reward signal during the optimization process, preventing the model from collapsing into specific modes and thereby maintaining diversity. Our comprehensive evaluation, combining qualitative analysis with quantitative metrics for both quality and diversity, reveals that D$^2$-Align achieves superior alignment with human preference.

Ziqi Wang, Xingzhou Lou, Meiqi Wu et al.

Group Relative Policy Optimization (GRPO) enhances LLM reasoning but often induces overconfidence, where incorrect responses yield lower perplexity than correct ones, degrading relative calibration as described by the Area Under the Curve (AUC). Existing approaches either yield limited improvements in calibration or sacrifice gains in reasoning accuracy. We first prove that this degradation in GRPO-style algorithms stems from their uncertainty-agnostic advantage estimation, which inevitably misaligns optimization gradients with calibration. This leads to improved accuracy at the expense of degraded calibration. We then propose Calibration-Aware Policy Optimization (CAPO). It adopts a logistic AUC surrogate loss that is theoretically consistent and admits regret bound, enabling uncertainty-aware advantage estimation. By further incorporating a noise masking mechanism, CAPO achieves stable learning dynamics that jointly optimize calibration and accuracy. Experiments on multiple mathematical reasoning benchmarks show that CAPO-1.5B significantly improves calibration by up to 15% while achieving accuracy comparable to or better than GRPO, and further boosts accuracy on downstream inference-time scaling tasks by up to 5%. Moreover, when allowed to abstain under low-confidence conditions, CAPO achieves a Pareto-optimal precision-coverage trade-off, highlighting its practical value for hallucination mitigation.

Xinran Ling, Chen Zhu, Meiqi Wu et al.

Video generation has advanced rapidly, improving evaluation methods, yet assessing video's motion remains a major challenge. Specifically, there are two key issues: 1) current motion metrics do not fully align with human perceptions; 2) the existing motion prompts are limited. Based on these findings, we introduce VMBench--a comprehensive Video Motion Benchmark that has perception-aligned motion metrics and features the most diverse types of motion. VMBench has several appealing properties: 1) Perception-Driven Motion Evaluation Metrics, we identify five dimensions based on human perception in motion video assessment and develop fine-grained evaluation metrics, providing deeper insights into models' strengths and weaknesses in motion quality. 2) Meta-Guided Motion Prompt Generation, a structured method that extracts meta-information, generates diverse motion prompts with LLMs, and refines them through human-AI validation, resulting in a multi-level prompt library covering six key dynamic scene dimensions. 3) Human-Aligned Validation Mechanism, we provide human preference annotations to validate our benchmarks, with our metrics achieving an average 35.3% improvement in Spearman's correlation over baseline methods. This is the first time that the quality of motion in videos has been evaluated from the perspective of human perception alignment. Additionally, we will soon release VMBench at https://github.com/GD-AIGC/VMBench, setting a new standard for evaluating and advancing motion generation models.

Meiqi Wu, Zhixin Cai, Fufangchen Zhao et al.

Video--based world models have emerged along two dominant paradigms: video generation and 3D reconstruction. However, existing evaluation benchmarks either focus narrowly on visual fidelity and text--video alignment for generative models, or rely on static 3D reconstruction metrics that fundamentally neglect temporal dynamics. We argue that the future of world modeling lies in 4D generation, which jointly models spatial structure and temporal evolution. In this paradigm, the core capability is interactive response: the ability to faithfully reflect how interaction actions drive state transitions across space and time. Yet no existing benchmark systematically evaluates this critical dimension. To address this gap, we propose Omni--WorldBench, a comprehensive benchmark specifically designed to evaluate the interactive response capabilities of world models in 4D settings. Omni--WorldBench comprises two key components: Omni--WorldSuite, a systematic prompt suite spanning diverse interaction levels and scene types; and Omni--Metrics, an agent-based evaluation framework that quantifies world modeling capabilities by measuring the causal impact of interaction actions on both final outcomes and intermediate state evolution trajectories. We conduct extensive evaluations of 18 representative world models across multiple paradigms. Our analysis reveals critical limitations of current world models in interactive response, providing actionable insights for future research. Omni-WorldBench will be publicly released to foster progress in interactive 4D world modeling.

Meiqi Wu, Kaiqi Huang, Yuanqiang Cai et al.

Air-writing is a challenging task that combines the fields of computer vision and natural language processing, offering an intuitive and natural approach for human-computer interaction. However, current air-writing solutions face two primary challenges: (1) their dependency on complex sensors (e.g., Radar, EEGs and others) for capturing precise handwritten trajectories, and (2) the absence of a video-based air-writing dataset that covers a comprehensive vocabulary range. These limitations impede their practicality in various real-world scenarios, including the use on devices like iPhones and laptops. To tackle these challenges, we present the groundbreaking air-writing Chinese character video dataset (AWCV-100K-UCAS2024), serving as a pioneering benchmark for video-based air-writing. This dataset captures handwritten trajectories in various real-world scenarios using commonly accessible RGB cameras, eliminating the need for complex sensors. AWCV-100K-UCAS2024 includes 8.8 million video frames, encompassing the complete set of 3,755 characters from the GB2312-80 level-1 set (GB1). Furthermore, we introduce our baseline approach, the video-based character recognizer (VCRec). VCRec adeptly extracts fingertip features from sparse visual cues and employs a spatio-temporal sequence module for analysis. Experimental results showcase the superior performance of VCRec compared to existing models in recognizing air-written characters, both quantitatively and qualitatively. This breakthrough paves the way for enhanced human-computer interaction in real-world contexts. Moreover, our approach leverages affordable RGB cameras, enabling its applicability in a diverse range of scenarios. The code and data examples will be made public at https://github.com/wmeiqi/AWCV.

Chubin Chen, Jiashu Zhu, Xiaokun Feng et al.

Classifier-free Guidance (CFG) is a widely used technique in modern diffusion models for enhancing sample quality and prompt adherence. However, through an empirical analysis on Gaussian mixture modeling with a closed-form solution, we observe a discrepancy between the suboptimal results produced by CFG and the ground truth. The model's excessive reliance on these suboptimal predictions often leads to semantic incoherence and low-quality outputs. To address this issue, we first empirically demonstrate that the model's suboptimal predictions can be effectively refined using sub-networks of the model itself. Building on this insight, we propose S^2-Guidance, a novel method that leverages stochastic block-dropping during the forward process to construct stochastic sub-networks, effectively guiding the model away from potential low-quality predictions and toward high-quality outputs. Extensive qualitative and quantitative experiments on text-to-image and text-to-video generation tasks demonstrate that S^2-Guidance delivers superior performance, consistently surpassing CFG and other advanced guidance strategies. Our code will be released.

Meiqi Wu, Yaxuan Kang, Xuchen Li et al.

The Drawing Projection Test (DPT) is an essential tool in art therapy, allowing psychologists to assess participants' mental states through their sketches. Specifically, through sketches with the theme of "a person picking an apple from a tree (PPAT)", it can be revealed whether the participants are in mental states such as depression. Compared with scales, the DPT can enrich psychologists' understanding of an individual's mental state. However, the interpretation of the PPAT is laborious and depends on the experience of the psychologists. To address this issue, we propose an effective identification method to support psychologists in conducting a large-scale automatic DPT. Unlike traditional sketch recognition, DPT more focus on the overall evaluation of the sketches, such as color usage and space utilization. Moreover, PPAT imposes a time limit and prohibits verbal reminders, resulting in low drawing accuracy and a lack of detailed depiction. To address these challenges, we propose the following efforts: (1) Providing an experimental environment for automated analysis of PPAT sketches for depression assessment; (2) Offering a Visual-Semantic depression assessment based on LLM (VS-LLM) method; (3) Experimental results demonstrate that our method improves by 17.6% compared to the psychologist assessment method. We anticipate that this work will contribute to the research in mental state assessment based on PPAT sketches' elements recognition. Our datasets and codes are available at https://github.com/wmeiqi/VS-LLM.

Meiqi Wu, Bingze Song, Ruimin Lin et al.

Video generation models have achieved notable progress in static scenarios, yet their performance in motion video generation remains limited, with quality degrading under drastic dynamic changes. This is due to noise disrupting temporal coherence and increasing the difficulty of learning dynamic regions. {Unfortunately, existing diffusion models rely on static loss for all scenarios, constraining their ability to capture complex dynamics.} To address this issue, we introduce Latent Temporal Discrepancy (LTD) as a motion prior to guide loss weighting. LTD measures frame-to-frame variation in the latent space, assigning larger penalties to regions with higher discrepancy while maintaining regular optimization for stable regions. This motion-aware strategy stabilizes training and enables the model to better reconstruct high-frequency dynamics. Extensive experiments on the general benchmark VBench and the motion-focused VMBench show consistent gains, with our method outperforming strong baselines by 3.31% on VBench and 3.58% on VMBench, achieving significant improvements in motion quality.

Chen Zhu, Jiashu Zhu, Yanxun Li et al.

With the rapid advancement of video generation techniques, evaluating and auditing generated videos has become increasingly crucial. Existing approaches typically offer coarse video quality scores, lacking detailed localization and categorization of specific artifacts. In this work, we introduce a comprehensive evaluation protocol focusing on three key aspects affecting human perception: Appearance, Motion, and Camera. We define these axes through a taxonomy of 10 prevalent artifact categories reflecting common generative failures observed in video generation. To enable robust artifact detection and categorization, we introduce GenVID, a large-scale dataset of 80k videos generated by various state-of-the-art video generation models, each carefully annotated for the defined artifact categories. Leveraging GenVID, we develop DVAR, a Dense Video Artifact Recognition framework for fine-grained identification and classification of generative artifacts. Extensive experiments show that our approach significantly improves artifact detection accuracy and enables effective filtering of low-quality content.

Xuchen Li, Xiaokun Feng, Shiyu Hu et al.

Visual Language Tracking (VLT) enhances single object tracking (SOT) by integrating natural language descriptions from a video, for the precise tracking of a specified object. By leveraging high-level semantic information, VLT guides object tracking, alleviating the constraints associated with relying on a visual modality. Nevertheless, most VLT benchmarks are annotated in a single granularity and lack a coherent semantic framework to provide scientific guidance. Moreover, coordinating human annotators for high-quality annotations is laborious and time-consuming. To address these challenges, we introduce DTLLM-VLT, which automatically generates extensive and multi-granularity text to enhance environmental diversity. (1) DTLLM-VLT generates scientific and multi-granularity text descriptions using a cohesive prompt framework. Its succinct and highly adaptable design allows seamless integration into various visual tracking benchmarks. (2) We select three prominent benchmarks to deploy our approach: short-term tracking, long-term tracking, and global instance tracking. We offer four granularity combinations for these benchmarks, considering the extent and density of semantic information, thereby showcasing the practicality and versatility of DTLLM-VLT. (3) We conduct comparative experiments on VLT benchmarks with different text granularities, evaluating and analyzing the impact of diverse text on tracking performance. Conclusionally, this work leverages LLM to provide multi-granularity semantic information for VLT task from efficient and diverse perspectives, enabling fine-grained evaluation of multi-modal trackers. In the future, we believe this work can be extended to more datasets to support vision datasets understanding.

Xuchen Li, Shiyu Hu, Xiaokun Feng et al.

Vision-language tracking (VLT) extends traditional single object tracking by incorporating textual information, providing semantic guidance to enhance tracking performance under challenging conditions like fast motion and deformations. However, current VLT trackers often underperform compared to single-modality methods on multiple benchmarks, with semantic information sometimes becoming a "distraction." To address this, we propose VLTVerse, the first fine-grained evaluation framework for VLT trackers that comprehensively considers multiple challenge factors and diverse semantic information, hoping to reveal the role of language in VLT. Our contributions include: (1) VLTVerse introduces 10 sequence-level challenge labels and 6 types of multi-granularity semantic information, creating a flexible and multi-dimensional evaluation space for VLT; (2) leveraging 60 subspaces formed by combinations of challenge factors and semantic types, we conduct systematic fine-grained evaluations of three mainstream SOTA VLT trackers, uncovering their performance bottlenecks across complex scenarios and offering a novel perspective on VLT evaluation; (3) through decoupled analysis of experimental results, we examine the impact of various semantic types on specific challenge factors in relation to different algorithms, providing essential guidance for enhancing VLT across data, evaluation, and algorithmic dimensions. The VLTVerse, toolkit, and results will be available at \url{http://metaverse.aitestunion.com}.

Fangyuan Mao, Aiming Hao, Jintao Chen et al.

Visual effects (VFX) are essential visual enhancements fundamental to modern cinematic production. Although video generation models offer cost-efficient solutions for VFX production, current methods are constrained by per-effect LoRA training, which limits generation to single effects. This fundamental limitation impedes applications that require spatially controllable composite effects, i.e., the concurrent generation of multiple effects at designated locations. However, integrating diverse effects into a unified framework faces major challenges: interference from effect variations and spatial uncontrollability during multi-VFX joint training. To tackle these challenges, we propose Omni-Effects, a first unified framework capable of generating prompt-guided effects and spatially controllable composite effects. The core of our framework comprises two key innovations: (1) LoRA-based Mixture of Experts (LoRA-MoE), which employs a group of expert LoRAs, integrating diverse effects within a unified model while effectively mitigating cross-task interference. (2) Spatial-Aware Prompt (SAP) incorporates spatial mask information into the text token, enabling precise spatial control. Furthermore, we introduce an Independent-Information Flow (IIF) module integrated within the SAP, isolating the control signals corresponding to individual effects to prevent any unwanted blending. To facilitate this research, we construct a comprehensive VFX dataset Omni-VFX via a novel data collection pipeline combining image editing and First-Last Frame-to-Video (FLF2V) synthesis, and introduce a dedicated VFX evaluation framework for validating model performance. Extensive experiments demonstrate that Omni-Effects achieves precise spatial control and diverse effect generation, enabling users to specify both the category and location of desired effects.

Meiqi Wu, Jiashu Zhu, Xiaokun Feng et al.

Video generation models have achieved remarkable progress, particularly excelling in realistic scenarios; however, their performance degrades notably in imaginative scenarios. These prompts often involve rarely co-occurring concepts with long-distance semantic relationships, falling outside training distributions. Existing methods typically apply test-time scaling for improving video quality, but their fixed search spaces and static reward designs limit adaptability to imaginative scenarios. To fill this gap, we propose ImagerySearch, a prompt-guided adaptive test-time search strategy that dynamically adjusts both the inference search space and reward function according to semantic relationships in the prompt. This enables more coherent and visually plausible videos in challenging imaginative settings. To evaluate progress in this direction, we introduce LDT-Bench, the first dedicated benchmark for long-distance semantic prompts, consisting of 2,839 diverse concept pairs and an automated protocol for assessing creative generation capabilities. Extensive experiments show that ImagerySearch consistently outperforms strong video generation baselines and existing test-time scaling approaches on LDT-Bench, and achieves competitive improvements on VBench, demonstrating its effectiveness across diverse prompt types. We will release LDT-Bench and code to facilitate future research on imaginative video generation.

Fangyuan Mao, Shuo Wang, Jilin Mei et al.

Joint RGB-infrared perception is essential for achieving robustness under diverse weather and illumination conditions. Although foundation models excel within single modalities, they suffer from substantial cross-modal degradation, an issue we attribute to a pattern shortcut, i.e., a modal bias that prioritizes superficial sensor patterns over underlying semantics. To address this problem, we introduce UNIV, a Unified foundation model for Infrared and Visible modalities. At the core of UNIV lies Patch Cross-modal Contrastive Learning (PCCL), a self-supervised contrastive learning strategy that constructs a unified cross-modal feature space. PCCL employs a frozen pre-trained model to sample pseudo patch pairs based on semantic similarity, and aligns infrared-visible representations by attracting semantically related pairs while repelling unrelated ones. This process simultaneously enhances cross-modal alignment and inter-class semantic separability, guiding the model to focus on semantic structure rather than falling into pattern shortcuts. To further enable cross-modal learning, we introduce MVIP, the most comprehensive visible-infrared benchmark to date, containing 98,992 precisely aligned image pairs across diverse scenes. Extensive experiments demonstrate UNIV's superior performance on infrared tasks (+1.7 mIoU for semantic segmentation and +0.7 mAP for detection), while maintaining competitive accuracy on RGB tasks.