Yunfei Liu

Yunfei Liu, Hao Wang, Yuhang Qi et al.

High-fidelity computational fluid dynamics is essential for aerospace design, but engineering-scale simulations of practical three-dimensional aircraft remain computationally expensive. Learning-based flow-field initialization can improve efficiency by reducing the numerical distance between the initial and converged solutions, yet existing deep learning approaches remain difficult to scale to large three-dimensional aircraft flows with multiscale regional heterogeneity. Most prior studies therefore focus on two-dimensional problems, surface quantities, integral aerodynamic coefficients, or simplified three-dimensional cases with limited grid resolution.Here we propose MHLF, a multigrid-hierarchical learning framework for accelerating engineering-scale aircraft flow simulations while preserving high-fidelity numerical accuracy. MHLF combines a topologically consistent geometric multigrid representation with a hierarchical strategy that captures regional flow heterogeneity during both prediction and subsequent CFD correction. Across three engineering-scale aircraft cases spanning Mach 0.15 to 6.0 and covering subsonic, transonic and supersonic regimes, MHLF accelerates convergence without sacrificing flow-field accuracy, achieving a 3 to 8 times efficiency improvement over conventional initialization. These results demonstrate practical full-flow-field prediction for large three-dimensional aircraft within the CFD domain and provide a foundation for data-driven acceleration of high-fidelity aircraft flow simulation.

Huarui He, Jie Wang, Yunfei Liu et al.

Single-step retrosynthesis is the cornerstone of retrosynthesis planning, which is a crucial task for computer-aided drug discovery. The goal of single-step retrosynthesis is to identify the possible reactants that lead to the synthesis of the target product in one reaction. By representing organic molecules as canonical strings, existing sequence-based retrosynthetic methods treat the product-to-reactant retrosynthesis as a sequence-to-sequence translation problem. However, most of them struggle to identify diverse chemical reactions for a desired product due to the deterministic inference, which contradicts the fact that many compounds can be synthesized through various reaction types with different sets of reactants. In this work, we aim to increase reaction diversity and generate various reactants using discrete latent variables. We propose a novel sequence-based approach, namely RetroDVCAE, which incorporates conditional variational autoencoders into single-step retrosynthesis and associates discrete latent variables with the generation process. Specifically, RetroDVCAE uses the Gumbel-Softmax distribution to approximate the categorical distribution over potential reactions and generates multiple sets of reactants with the variational decoder. Experiments demonstrate that RetroDVCAE outperforms state-of-the-art baselines on both benchmark dataset and homemade dataset. Both quantitative and qualitative results show that RetroDVCAE can model the multi-modal distribution over reaction types and produce diverse reactant candidates.

Zongji Wang, Yunfei Liu, Feng Lu

Intrinsic image decomposition is an important and long-standing computer vision problem. Given an input image, recovering the physical scene properties is ill-posed. Several physically motivated priors have been used to restrict the solution space of the optimization problem for intrinsic image decomposition. This work takes advantage of deep learning, and shows that it can solve this challenging computer vision problem with high efficiency. The focus lies in the feature encoding phase to extract discriminative features for different intrinsic layers from an input image. To achieve this goal, we explore the distinctive characteristics of different intrinsic components in the high dimensional feature embedding space. We define feature distribution divergence to efficiently separate the feature vectors of different intrinsic components. The feature distributions are also constrained to fit the real ones through a feature distribution consistency. In addition, a data refinement approach is provided to remove data inconsistency from the Sintel dataset, making it more suitable for intrinsic image decomposition. Our method is also extended to intrinsic video decomposition based on pixel-wise correspondences between adjacent frames. Experimental results indicate that our proposed network structure can outperform the existing state-of-the-art.

Ruicong Liu, Yiwei Bao, Mingjie Xu et al.

Deep neural networks have demonstrated superior performance on appearance-based gaze estimation tasks. However, due to variations in person, illuminations, and background, performance degrades dramatically when applying the model to a new domain. In this paper, we discover an interesting gaze jitter phenomenon in cross-domain gaze estimation, i.e., the gaze predictions of two similar images can be severely deviated in target domain. This is closely related to cross-domain gaze estimation tasks, but surprisingly, it has not been noticed yet previously. Therefore, we innovatively propose to utilize the gaze jitter to analyze and optimize the gaze domain adaptation task. We find that the high-frequency component (HFC) is an important factor that leads to jitter. Based on this discovery, we add high-frequency components to input images using the adversarial attack and employ contrastive learning to encourage the model to obtain similar representations between original and perturbed data, which reduces the impacts of HFC. We evaluate the proposed method on four cross-domain gaze estimation tasks, and experimental results demonstrate that it significantly reduces the gaze jitter and improves the gaze estimation performance in target domains.

Yunfei Liu, Lijian Lin, Fei Yu et al.

Audio-driven portrait animation aims to synthesize portrait videos that are conditioned by given audio. Animating high-fidelity and multimodal video portraits has a variety of applications. Previous methods have attempted to capture different motion modes and generate high-fidelity portrait videos by training different models or sampling signals from given videos. However, lacking correlation learning between lip-sync and other movements (e.g., head pose/eye blinking) usually leads to unnatural results. In this paper, we propose a unified system for multi-person, diverse, and high-fidelity talking portrait generation. Our method contains three stages, i.e., 1) Mapping-Once network with Dual Attentions (MODA) generates talking representation from given audio. In MODA, we design a dual-attention module to encode accurate mouth movements and diverse modalities. 2) Facial composer network generates dense and detailed face landmarks, and 3) temporal-guided renderer syntheses stable videos. Extensive evaluations demonstrate that the proposed system produces more natural and realistic video portraits compared to previous methods.

Mingfang Zhang, Yunfei Liu, Feng Lu

Appearance-based gaze estimation aims to predict the 3D eye gaze direction from a single image. While recent deep learning-based approaches have demonstrated excellent performance, they usually assume one calibrated face in each input image and cannot output multi-person gaze in real time. However, simultaneous gaze estimation for multiple people in the wild is necessary for real-world applications. In this paper, we propose the first one-stage end-to-end gaze estimation method, GazeOnce, which is capable of simultaneously predicting gaze directions for multiple faces (>10) in an image. In addition, we design a sophisticated data generation pipeline and propose a new dataset, MPSGaze, which contains full images of multiple people with 3D gaze ground truth. Experimental results demonstrate that our unified framework not only offers a faster speed, but also provides a lower gaze estimation error compared with state-of-the-art methods. This technique can be useful in real-time applications with multiple users.

Qisong Xiao, Xinhai Chen, Qinglin Wang et al.

Deep learning has emerged as a promising paradigm for spatio-temporal modeling of fluid dynamics. However, existing approaches often suffer from limited generalization to unseen flow conditions and typically require retraining when applied to new scenarios. In this paper, we present LLM4Fluid, a spatio-temporal prediction framework that leverages Large Language Models (LLMs) as generalizable neural solvers for fluid dynamics. The framework first compresses high-dimensional flow fields into a compact latent space via reduced-order modeling enhanced with a physics-informed disentanglement mechanism, effectively mitigating spatial feature entanglement while preserving essential flow structures. A pretrained LLM then serves as a temporal processor, autoregressively predicting the dynamics of physical sequences with time series prompts. To bridge the modality gap between prompts and physical sequences, which can otherwise degrade prediction accuracy, we propose a dedicated modality alignment strategy that resolves representational mismatch and stabilizes long-term prediction. Extensive experiments across diverse flow scenarios demonstrate that LLM4Fluid functions as a robust and generalizable neural solver without retraining, achieving state-of-the-art accuracy while exhibiting powerful zero-shot and in-context learning capabilities. Code and datasets are publicly available at https://github.com/qisongxiao/LLM4Fluid.

Liying Lu, Tianke Zhang, Yunfei Liu et al.

Given an arbitrary audio clip, audio-driven 3D facial animation aims to generate lifelike lip motions and facial expressions for a 3D head. Existing methods typically rely on training their models using limited public 3D datasets that contain a restricted number of audio-3D scan pairs. Consequently, their generalization capability remains limited. In this paper, we propose a novel method that leverages in-the-wild 2D talking-head videos to train our 3D facial animation model. The abundance of easily accessible 2D talking-head videos equips our model with a robust generalization capability. By combining these videos with existing 3D face reconstruction methods, our model excels in generating consistent and high-fidelity lip synchronization. Additionally, our model proficiently captures the speaking styles of different individuals, allowing it to generate 3D talking-heads with distinct personal styles. Extensive qualitative and quantitative experimental results demonstrate the superiority of our method.

Chaoqun Zhuang, Yunfei Liu, Sijia Wen et al.

Low-light hazy scenes commonly appear at dusk and early morning. The visual enhancement for low-light hazy images is an ill-posed problem. Even though numerous methods have been proposed for image dehazing and low-light enhancement respectively, simply integrating them cannot deliver pleasing results for this particular task. In this paper, we present a novel method to enhance visibility for low-light hazy scenarios. To handle this challenging task, we propose two key techniques, namely cross-consistency dehazing-enhancement framework and physically based simulation for low-light hazy dataset. Specifically, the framework is designed for enhancing visibility of the input image via fully utilizing the clues from different sub-tasks. The simulation is designed for generating the dataset with ground-truths by the proposed low-light hazy imaging model. The extensive experimental results show that the proposed method outperforms the SOTA solutions on different metrics including SSIM (9.19%) and PSNR(5.03%). In addition, we conduct a user study on real images to demonstrate the effectiveness and necessity of the proposed method by human visual perception.

Jintang Li, Jiawang Dan, Ruofan Wu et al.

Over the past few years, graph neural networks (GNNs) have become powerful and practical tools for learning on (static) graph-structure data. However, many real-world applications, such as social networks and e-commerce, involve temporal graphs where nodes and edges are dynamically evolving. Temporal graph neural networks (TGNNs) have progressively emerged as an extension of GNNs to address time-evolving graphs and have gradually become a trending research topic in both academics and industry. Advancing research and application in such an emerging field necessitates the development of new tools to compose TGNN models and unify their different schemes for dealing with temporal graphs. In this work, we introduce LasTGL, an industrial framework that integrates unified and extensible implementations of common temporal graph learning algorithms for various advanced tasks. The purpose of LasTGL is to provide the essential building blocks for solving temporal graph learning tasks, focusing on the guiding principles of user-friendliness and quick prototyping on which PyTorch is based. In particular, LasTGL provides comprehensive temporal graph datasets, TGNN models and utilities along with well-documented tutorials, making it suitable for both absolute beginners and expert deep learning practitioners alike.

Zhangxuan Gu, Zhengwen Zeng, Zhenyu Xu et al.

We present UI-Venus, a native UI agent that takes only screenshots as input based on a multimodal large language model. UI-Venus achieves SOTA performance on both UI grounding and navigation tasks using only several hundred thousand high-quality training samples through reinforcement finetune (RFT) based on Qwen2.5-VL. Specifically, the 7B and 72B variants of UI-Venus obtain 94.1% / 50.8% and 95.3% / 61.9% on the standard grounding benchmarks, i.e., Screenspot-V2 / Pro, surpassing the previous SOTA baselines including open-source GTA1 and closed-source UI-TARS-1.5. To show UI-Venus's summary and planing ability, we also evaluate it on the AndroidWorld, an online UI navigation arena, on which our 7B and 72B variants achieve 49.1% and 65.9% success rate, also beating existing models. To achieve this, we introduce carefully designed reward functions for both UI grounding and navigation tasks and corresponding efficient data cleaning strategies. To further boost navigation performance, we propose Self-Evolving Trajectory History Alignment & Sparse Action Enhancement that refine historical reasoning traces and balances the distribution of sparse but critical actions, leading to more coherent planning and better generalization in complex UI tasks. Our contributions include the publish of SOTA open-source UI agents, comprehensive data cleaning protocols and a novel self-evolving framework for improving navigation performance, which encourage further research and development in the community. Code is available at https://github.com/inclusionAI/UI-Venus.

Fa-Ting Hong, Yunfei Liu, Yu Li et al.

Audio-driven talking head synthesis strives to generate lifelike video portraits from provided audio. The diffusion model, recognized for its superior quality and robust generalization, has been explored for this task. However, establishing a robust correspondence between temporal audio cues and corresponding spatial facial expressions with diffusion models remains a significant challenge in talking head generation. To bridge this gap, we present DreamHead, a hierarchical diffusion framework that learns spatial-temporal correspondences in talking head synthesis without compromising the model's intrinsic quality and adaptability.~DreamHead learns to predict dense facial landmarks from audios as intermediate signals to model the spatial and temporal correspondences.~Specifically, a first hierarchy of audio-to-landmark diffusion is first designed to predict temporally smooth and accurate landmark sequences given audio sequence signals. Then, a second hierarchy of landmark-to-image diffusion is further proposed to produce spatially consistent facial portrait videos, by modeling spatial correspondences between the dense facial landmark and appearance. Extensive experiments show that proposed DreamHead can effectively learn spatial-temporal consistency with the designed hierarchical diffusion and produce high-fidelity audio-driven talking head videos for multiple identities.

Jiahao Wu, Yunfei Liu, Lijian Lin et al.

Reconstructing detailed 3D human meshes from a single in-the-wild image remains a fundamental challenge in computer vision. Existing SMPLX-based methods often suffer from slow inference, produce only coarse body poses, and exhibit misalignments or unnatural artifacts in fine-grained regions such as the face and hands. These issues make current approaches difficult to apply to downstream tasks. To address these challenges, we propose PEAR-a fast and robust framework for pixel-aligned expressive human mesh recovery. PEAR explicitly tackles three major limitations of existing methods: slow inference, inaccurate localization of fine-grained human pose details, and insufficient facial expression capture. Specifically, to enable real-time SMPLX parameter inference, we depart from prior designs that rely on high resolution inputs or multi-branch architectures. Instead, we adopt a clean and unified ViT-based model capable of recovering coarse 3D human geometry. To compensate for the loss of fine-grained details caused by this simplified architecture, we introduce pixel-level supervision to optimize the geometry, significantly improving the reconstruction accuracy of fine-grained human details. To make this approach practical, we further propose a modular data annotation strategy that enriches the training data and enhances the robustness of the model. Overall, PEAR is a preprocessing-free framework that can simultaneously infer EHM-s (SMPLX and scaled-FLAME) parameters at over 100 FPS. Extensive experiments on multiple benchmark datasets demonstrate that our method achieves substantial improvements in pose estimation accuracy compared to previous SMPLX-based approaches. Project page: https://wujh2001.github.io/PEAR

Lei Zhu, Lijian Lin, Ye Zhu et al.

Current audio-driven 3D head generation methods mainly focus on single-speaker scenarios, lacking natural, bidirectional listen-and-speak interaction. Achieving seamless conversational behavior, where speaking and listening states transition fluidly remains a key challenge. Existing 3D conversational avatar approaches rely on error-prone pseudo-3D labels that fail to capture fine-grained facial dynamics. To address these limitations, we introduce a novel two-stage framework MANGO, which leveraging pure image-level supervision by alternately training to mitigate the noise introduced by pseudo-3D labels, thereby achieving better alignment with real-world conversational behaviors. Specifically, in the first stage, a diffusion-based transformer with a dual-audio interaction module models natural 3D motion from multi-speaker audio. In the second stage, we use a fast 3D Gaussian Renderer to generate high-fidelity images and provide 2D-level photometric supervision for the 3D motions through alternate training. Additionally, we introduce MANGO-Dialog, a high-quality dataset with over 50 hours of aligned 2D-3D conversational data across 500+ identities. Extensive experiments demonstrate that our method achieves exceptional accuracy and realism in modeling two-person 3D dialogue motion, significantly advancing the fidelity and controllability of audio-driven talking heads.

Luqing Luo, Wenjin Gui, Yunfei Liu et al.

Deep understanding of electromagnetic signals is fundamental to dynamic spectrum management, intelligent transportation, autonomous driving and unmanned vehicle perception. The field faces challenges because electromagnetic signals differ greatly from text and images, showing high heterogeneity, strong background noise and complex joint time frequency structure, which prevents existing general models from direct use. Electromagnetic communication and sensing tasks are diverse, current methods lack cross task generalization and transfer efficiency, and the scarcity of large high quality datasets blocks the creation of a truly general multitask learning framework. To overcome these issue, we introduce EMind, an electromagnetic signals foundation model that bridges large scale pretraining and the unique nature of this modality. We build the first unified and largest standardized electromagnetic signal dataset covering multiple signal types and tasks. By exploiting the physical properties of electromagnetic signals, we devise a length adaptive multi-signal packing method and a hardware-aware training strategy that enable efficient use and representation learning from heterogeneous multi-source signals. Experiments show that EMind achieves strong performance and broad generalization across many downstream tasks, moving decisively from task specific models to a unified framework for electromagnetic intelligence. The code is available at: https://github.com/GabrielleTse/EMind.

Junming Chen, Yunfei Liu, Jianan Wang et al.

We propose DiffSHEG, a Diffusion-based approach for Speech-driven Holistic 3D Expression and Gesture generation with arbitrary length. While previous works focused on co-speech gesture or expression generation individually, the joint generation of synchronized expressions and gestures remains barely explored. To address this, our diffusion-based co-speech motion generation transformer enables uni-directional information flow from expression to gesture, facilitating improved matching of joint expression-gesture distributions. Furthermore, we introduce an outpainting-based sampling strategy for arbitrary long sequence generation in diffusion models, offering flexibility and computational efficiency. Our method provides a practical solution that produces high-quality synchronized expression and gesture generation driven by speech. Evaluated on two public datasets, our approach achieves state-of-the-art performance both quantitatively and qualitatively. Additionally, a user study confirms the superiority of DiffSHEG over prior approaches. By enabling the real-time generation of expressive and synchronized motions, DiffSHEG showcases its potential for various applications in the development of digital humans and embodied agents.

Dongbin Zhang, Yunfei Liu, Lijian Lin et al.

Reconstructing animatable and high-quality 3D head avatars from monocular videos, especially with realistic relighting, is a valuable task. However, the limited information from single-view input, combined with the complex head poses and facial movements, makes this challenging. Previous methods achieve real-time performance by combining 3D Gaussian Splatting with a parametric head model, but the resulting head quality suffers from inaccurate face tracking and limited expressiveness of the deformation model. These methods also fail to produce realistic effects under novel lighting conditions. To address these issues, we propose HRAvatar, a 3DGS-based method that reconstructs high-fidelity, relightable 3D head avatars. HRAvatar reduces tracking errors through end-to-end optimization and better captures individual facial deformations using learnable blendshapes and learnable linear blend skinning. Additionally, it decomposes head appearance into several physical properties and incorporates physically-based shading to account for environmental lighting. Extensive experiments demonstrate that HRAvatar not only reconstructs superior-quality heads but also achieves realistic visual effects under varying lighting conditions.

Yunfei Liu, Lei Zhu, Lijian Lin et al.

3D facial reconstruction from a single in-the-wild image is a crucial task in human-centered computer vision tasks. While existing methods can recover accurate facial shapes, there remains significant space for improvement in fine-grained expression capture. Current approaches struggle with irregular mouth shapes, exaggerated expressions, and asymmetrical facial movements. We present TEASER (Token EnhAnced Spatial modeling for Expressions Reconstruction), which addresses these challenges and enhances 3D facial geometry performance. TEASER tackles two main limitations of existing methods: insufficient photometric loss for self-reconstruction and inaccurate localization of subtle expressions. We introduce a multi-scale tokenizer to extract facial appearance information. Combined with a neural renderer, these tokens provide precise geometric guidance for expression reconstruction. Furthermore, TEASER incorporates a pose-dependent landmark loss to further improve geometric performances. Our approach not only significantly enhances expression reconstruction quality but also offers interpretable tokens suitable for various downstream applications, such as photorealistic facial video driving, expression transfer, and identity swapping. Quantitative and qualitative experimental results across multiple datasets demonstrate that TEASER achieves state-of-the-art performance in precise expression reconstruction.

Dongbin Zhang, Yunfei Liu, Lijian Lin et al.

Reconstructing a high-quality, animatable 3D human avatar with expressive facial and hand motions from a single image has gained significant attention due to its broad application potential. 3D human avatar reconstruction typically requires multi-view or monocular videos and training on individual IDs, which is both complex and time-consuming. Furthermore, limited by SMPLX's expressiveness, these methods often focus on body motion but struggle with facial expressions. To address these challenges, we first introduce an expressive human model (EHM) to enhance facial expression capabilities and develop an accurate tracking method. Based on this template model, we propose GUAVA, the first framework for fast animatable upper-body 3D Gaussian avatar reconstruction. We leverage inverse texture mapping and projection sampling techniques to infer Ubody (upper-body) Gaussians from a single image. The rendered images are refined through a neural refiner. Experimental results demonstrate that GUAVA significantly outperforms previous methods in rendering quality and offers significant speed improvements, with reconstruction times in the sub-second range (0.1s), and supports real-time animation and rendering.

Maomao Li, Lijian Lin, Yunfei Liu et al.

This paper presents Qffusion, a dual-frame-guided framework for portrait video editing. Specifically, we consider a design principle of ``animation for editing'', and train Qffusion as a general animation framework from two still reference images while we can use it for portrait video editing easily by applying modified start and end frames as references during inference. Leveraging the powerful generative power of Stable Diffusion, we propose a Quadrant-grid Arrangement (QGA) scheme for latent re-arrangement, which arranges the latent codes of two reference images and that of four facial conditions into a four-grid fashion, separately. Then, we fuse features of these two modalities and use self-attention for both appearance and temporal learning, where representations at different times are jointly modeled under QGA. Our Qffusion can achieve stable video editing without additional networks or complex training stages, where only the input format of Stable Diffusion is modified. Further, we propose a Quadrant-grid Propagation (QGP) inference strategy, which enjoys a unique advantage on stable arbitrary-length video generation by processing reference and condition frames recursively. Through extensive experiments, Qffusion consistently outperforms state-of-the-art techniques on portrait video editing. Project page: https://qffusion.github.io/page/.

Maomao Li, Yu Li, Yunfei Liu et al.

Editability and fidelity are two essential demands for text-driven image editing, which expects that the editing area should align with the target prompt and the rest remain unchanged separately. The current cutting-edge editing methods usually obey an "inversion-then-editing" pipeline, where the input image is inverted to an approximate Gaussian noise ${z}_T$, based on which a sampling process is conducted using the target prompt. Nevertheless, we argue that it is not a good choice to use a near-Gaussian noise as a pivot for further editing since it would bring plentiful fidelity errors. We verify this by a pilot analysis, discovering that intermediate-inverted latents can achieve a better trade-off between editability and fidelity than the fully-inverted ${z}_T$. Based on this, we propose a novel zero-shot editing paradigm dubbed ZZEdit, which first locates a qualified intermediate-inverted latent marked as ${z}_p$ as a better editing pivot, which is sufficient-for-editing while structure-preserving. Then, a ZigZag process is designed to execute denoising and inversion alternately, which progressively inject target guidance to ${z}_p$ while preserving the structure information of $p$ step. Afterwards, to achieve the same step number of inversion and denoising, we execute a pure sampling process under the target prompt. Essentially, our ZZEdit performs iterative manifold constraint between the manifold of $M_{p}$ and $M_{p-1}$, leading to fewer fidelity errors. Extensive experiments highlight the effectiveness of ZZEdit in diverse image editing scenarios compared with the "inversion-then-editing" pipeline.

Maomao Li, Yunfei Liu, Yu Li

Image-driven video editing aims to propagate edit contents from the modified first frame to the rest frames. The existing methods usually invert the source video to noise using a pre-trained image-to-video (I2V) model and then guide the sampling process using the edited first frame. Generally, a popular choice for maintaining motion and layout from the source video is intervening in the denoising process by injecting attention during reconstruction. However, such injection often leads to unsatisfactory results, where excessive injection leads to conflicting semantics from the source video while insufficient injection brings limited source representation. Recognizing this, we propose an Editing-awaRE (REE) injection method to modulate injection intensity of each token. Specifically, we first compute the pixel difference between the source and edited first frame to form a corresponding editing mask. Next, we track the editing area throughout the entire video by using optical flow to warp the first-frame mask. Then, editing-aware feature injection intensity for each token is generated accordingly, where injection is not conducted on editing areas. Building upon REE injection, we further propose a zero-shot image-driven video editing framework with recent-emerging rectified-Flow models, dubbed FREE-Edit. Without fine-tuning or training, our FREE-Edit demonstrates effectiveness in various image-driven video editing scenarios, showing its capability to produce higher-quality outputs compared with existing techniques. Project page: https://free-edit.github.io/page/.

Bin Liu, Yunfei Liu, Ziru Xu et al.

Online advertising systems typically use a cascaded architecture to manage massive requests and candidate volumes, where the ranking stages allocate traffic based on eCPM (predicted CTR $\times$ Bid). With the increasing popularity of auto-bidding strategies, the inconsistency between the computationally sensitive retrieval stage and the ranking stages becomes more pronounced, as the former cannot access precise, real-time bids for the vast ad corpus. This discrepancy leads to sub-optimal platform revenue and advertiser outcomes. To tackle this problem, we propose Bidding-Aware Retrieval (BAR), a model-based retrieval framework that addresses multi-stage inconsistency by incorporating ad bid value into the retrieval scoring function. The core innovation is Bidding-Aware Modeling, incorporating bid signals through monotonicity-constrained learning and multi-task distillation to ensure economically coherent representations, while Asynchronous Near-Line Inference enables real-time updates to the embedding for market responsiveness. Furthermore, the Task-Attentive Refinement module selectively enhances feature interactions to disentangle user interest and commercial value signals. Extensive offline experiments and full-scale deployment across Alibaba's display advertising platform validated BAR's efficacy: 4.32% platform revenue increase with 22.2% impression lift for positively-operated advertisements.

Xiangyang Luo, Ye Zhu, Yunfei Liu et al.

Video face swapping aims to address two primary challenges: effectively transferring the source identity to the target video and accurately preserving the dynamic attributes of the target face, such as head poses, facial expressions, lip-sync, \etc. Existing methods mainly focus on achieving high-quality identity transfer but often fall short in maintaining the dynamic attributes of the target face, leading to inconsistent results. We attribute this issue to the inherent coupling of facial appearance and motion in videos. To address this, we propose CanonSwap, a novel video face-swapping framework that decouples motion information from appearance information. Specifically, CanonSwap first eliminates motion-related information, enabling identity modification within a unified canonical space. Subsequently, the swapped feature is reintegrated into the original video space, ensuring the preservation of the target face's dynamic attributes. To further achieve precise identity transfer with minimal artifacts and enhanced realism, we design a Partial Identity Modulation module that adaptively integrates source identity features using a spatial mask to restrict modifications to facial regions. Additionally, we introduce several fine-grained synchronization metrics to comprehensively evaluate the performance of video face swapping methods. Extensive experiments demonstrate that our method significantly outperforms existing approaches in terms of visual quality, temporal consistency, and identity preservation. Our project page are publicly available at https://luoxyhappy.github.io/CanonSwap/.

Runzhen Liu, Qinjie Lin, Yunfei Liu et al.

Video dubbing aims to synthesize realistic, lip-synced videos from a reference video and a driving audio signal. Although existing methods can accurately generate mouth shapes driven by audio, they often fail to preserve identity-specific features, largely because they do not effectively capture the nuanced interplay between audio cues and the visual attributes of reference identity . As a result, the generated outputs frequently lack fidelity in reproducing the unique textural and structural details of the reference identity. To address these limitations, we propose IPTalker, a novel and robust framework for video dubbing that achieves seamless alignment between driving audio and reference identity while ensuring both lip-sync accuracy and high-fidelity identity preservation. At the core of IPTalker is a transformer-based alignment mechanism designed to dynamically capture and model the correspondence between audio features and reference images, thereby enabling precise, identity-aware audio-visual integration. Building on this alignment, a motion warping strategy further refines the results by spatially deforming reference images to match the target audio-driven configuration. A dedicated refinement process then mitigates occlusion artifacts and enhances the preservation of fine-grained textures, such as mouth details and skin features. Extensive qualitative and quantitative evaluations demonstrate that IPTalker consistently outperforms existing approaches in terms of realism, lip synchronization, and identity retention, establishing a new state of the art for high-quality, identity-consistent video dubbing.

Yunfei Liu, Jintang Li, Yuehe Chen et al.

Graph clustering, a fundamental and challenging task in graph mining, aims to classify nodes in a graph into several disjoint clusters. In recent years, graph contrastive learning (GCL) has emerged as a dominant line of research in graph clustering and advances the new state-of-the-art. However, GCL-based methods heavily rely on graph augmentations and contrastive schemes, which may potentially introduce challenges such as semantic drift and scalability issues. Another promising line of research involves the adoption of modularity maximization, a popular and effective measure for community detection, as the guiding principle for clustering tasks. Despite the recent progress, the underlying mechanism of modularity maximization is still not well understood. In this work, we dig into the hidden success of modularity maximization for graph clustering. Our analysis reveals the strong connections between modularity maximization and graph contrastive learning, where positive and negative examples are naturally defined by modularity. In light of our results, we propose a community-aware graph clustering framework, coined MAGI, which leverages modularity maximization as a contrastive pretext task to effectively uncover the underlying information of communities in graphs, while avoiding the problem of semantic drift. Extensive experiments on multiple graph datasets verify the effectiveness of MAGI in terms of scalability and clustering performance compared to state-of-the-art graph clustering methods. Notably, MAGI easily scales a sufficiently large graph with 100M nodes while outperforming strong baselines.

Weizhen Liu, Ao Li, Ze Wu et al.

Hierarchical leaf vein segmentation is a crucial but under-explored task in agricultural sciences, where analysis of the hierarchical structure of plant leaf venation can contribute to plant breeding. While current segmentation techniques rely on data-driven models, there is no publicly available dataset specifically designed for hierarchical leaf vein segmentation. To address this gap, we introduce the HierArchical Leaf Vein Segmentation (HALVS) dataset, the first public hierarchical leaf vein segmentation dataset. HALVS comprises 5,057 real-scanned high-resolution leaf images collected from three plant species: soybean, sweet cherry, and London planetree. It also includes human-annotated ground truth for three orders of leaf veins, with a total labeling effort of 83.8 person-days. Based on HALVS, we further develop a label-efficient learning paradigm that leverages partial label information, i.e. missing annotations for tertiary veins. Empirical studies are performed on HALVS, revealing new observations, challenges, and research directions on leaf vein segmentation.

Xuangeng Chu, Yu Li, Ailing Zeng et al.

Head avatar reconstruction, crucial for applications in virtual reality, online meetings, gaming, and film industries, has garnered substantial attention within the computer vision community. The fundamental objective of this field is to faithfully recreate the head avatar and precisely control expressions and postures. Existing methods, categorized into 2D-based warping, mesh-based, and neural rendering approaches, present challenges in maintaining multi-view consistency, incorporating non-facial information, and generalizing to new identities. In this paper, we propose a framework named GPAvatar that reconstructs 3D head avatars from one or several images in a single forward pass. The key idea of this work is to introduce a dynamic point-based expression field driven by a point cloud to precisely and effectively capture expressions. Furthermore, we use a Multi Tri-planes Attention (MTA) fusion module in the tri-planes canonical field to leverage information from multiple input images. The proposed method achieves faithful identity reconstruction, precise expression control, and multi-view consistency, demonstrating promising results for free-viewpoint rendering and novel view synthesis.

Maomao Li, Yu Li, Tianyu Yang et al.

This paper presents a video inversion approach for zero-shot video editing, which models the input video with low-rank representation during the inversion process. The existing video editing methods usually apply the typical 2D DDIM inversion or naive spatial-temporal DDIM inversion before editing, which leverages time-varying representation for each frame to derive noisy latent. Unlike most existing approaches, we propose a Spatial-Temporal Expectation-Maximization (STEM) inversion, which formulates the dense video feature under an expectation-maximization manner and iteratively estimates a more compact basis set to represent the whole video. Each frame applies the fixed and global representation for inversion, which is more friendly for temporal consistency during reconstruction and editing. Extensive qualitative and quantitative experiments demonstrate that our STEM inversion can achieve consistent improvement on two state-of-the-art video editing methods. Project page: https://stem-inv.github.io/page/.

Zongji Wang, Yunfei Liu, Feng Lu

In the area of 3D shape analysis, the geometric properties of a shape have long been studied. Instead of directly extracting representative features using expert-designed descriptors or end-to-end deep neural networks, this paper is dedicated to discovering distinctive information from the shape formation process. Concretely, a spherical point cloud served as the template is progressively deformed to fit the target shape in a coarse-to-fine manner. During the shape formation process, several checkpoints are inserted to facilitate recording and investigating the intermediate stages. For each stage, the offset field is evaluated as a stage-aware description. The summation of the offsets throughout the shape formation process can completely define the target shape in terms of geometry. In this perspective, one can derive the point-wise shape correspondence from the template inexpensively, which benefits various graphic applications. In this paper, the Progressive Deformation-based Auto-Encoder (PDAE) is proposed to learn the stage-aware description through a coarse-to-fine shape fitting task. Experimental results show that the proposed PDAE has the ability to reconstruct 3D shapes with high fidelity, and consistent topology is preserved in the multi-stage deformation process. Additional applications based on the stage-aware description are performed, demonstrating its universality.

Yunfei Liu, Haofei Wang, Yang Yue et al.

Unsupervised image-to-image translation aims to learn the mapping between two visual domains with unpaired samples. Existing works focus on disentangling domain-invariant content code and domain-specific style code individually for multimodal purposes. However, less attention has been paid to interpreting and manipulating the translated image. In this paper, we propose to separate the content code and style code simultaneously in a unified framework. Based on the correlation between the latent features and the high-level domain-invariant tasks, the proposed framework demonstrates superior performance in multimodal translation, interpretability and manipulation of the translated image. Experimental results show that the proposed approach outperforms the existing unsupervised image translation methods in terms of visual quality and diversity.

Yunfei Liu, Ruicong Liu, Haofei Wang et al.

Deep neural networks have significantly improved appearance-based gaze estimation accuracy. However, it still suffers from unsatisfactory performance when generalizing the trained model to new domains, e.g., unseen environments or persons. In this paper, we propose a plug-and-play gaze adaptation framework (PnP-GA), which is an ensemble of networks that learn collaboratively with the guidance of outliers. Since our proposed framework does not require ground-truth labels in the target domain, the existing gaze estimation networks can be directly plugged into PnP-GA and generalize the algorithms to new domains. We test PnP-GA on four gaze domain adaptation tasks, ETH-to-MPII, ETH-to-EyeDiap, Gaze360-to-MPII, and Gaze360-to-EyeDiap. The experimental results demonstrate that the PnP-GA framework achieves considerable performance improvements of 36.9%, 31.6%, 19.4%, and 11.8% over the baseline system. The proposed framework also outperforms the state-of-the-art domain adaptation approaches on gaze domain adaptation tasks.

Mingjie Xu, Haofei Wang, Yunfei Liu et al.

Appearance-based gaze estimation has achieved significant improvement by using deep learning. However, many deep learning-based methods suffer from the vulnerability property, i.e., perturbing the raw image using noise confuses the gaze estimation models. Although the perturbed image visually looks similar to the original image, the gaze estimation models output the wrong gaze direction. In this paper, we investigate the vulnerability of appearance-based gaze estimation. To our knowledge, this is the first time that the vulnerability of gaze estimation to be found. We systematically characterized the vulnerability property from multiple aspects, the pixel-based adversarial attack, the patch-based adversarial attack and the defense strategy. Our experimental results demonstrate that the CA-Net shows superior performance against attack among the four popular appearance-based gaze estimation networks, Full-Face, Gaze-Net, CA-Net and RT-GENE. This study draws the attention of researchers in the appearance-based gaze estimation community to defense from adversarial attacks.

Yunfei Liu, Chaoqun Zhuang, Feng Lu

Anomaly detection from a single image is challenging since anomaly data is always rare and can be with highly unpredictable types. With only anomaly-free data available, most existing methods train an AutoEncoder to reconstruct the input image and find the difference between the input and output to identify the anomalous region. However, such methods face a potential problem - a coarse reconstruction generates extra image differences while a high-fidelity one may draw in the anomaly. In this paper, we solve this contradiction by proposing a two-stage approach, which generates high-fidelity yet anomaly-free reconstructions. Our Unsupervised Two-stage Anomaly Detection (UTAD) relies on two technical components, namely the Impression Extractor (IE-Net) and the Expert-Net. The IE-Net and Expert-Net accomplish the two-stage anomaly-free image reconstruction task while they also generate intuitive intermediate results, making the whole UTAD interpretable. Extensive experiments show that our method outperforms state-of-the-arts on four anomaly detection datasets with different types of real-world objects and textures.

Yiwei Bao, Yihua Cheng, Yunfei Liu et al.

Recently, many multi-stream gaze estimation methods have been proposed. They estimate gaze from eye and face appearances and achieve reasonable accuracy. However, most of the methods simply concatenate the features extracted from eye and face appearance. The feature fusion process has been ignored. In this paper, we propose a novel Adaptive Feature Fusion Network (AFF-Net), which performs gaze tracking task in mobile tablets. We stack two-eye feature maps and utilize Squeeze-and-Excitation layers to adaptively fuse two-eye features according to their similarity on appearance. Meanwhile, we also propose Adaptive Group Normalization to recalibrate eye features with the guidance of facial feature. Extensive experiments on both GazeCapture and MPIIFaceGaze datasets demonstrate consistently superior performance of the proposed method.

Yunfei Liu, Yang Yang, Xianyu Chen et al.

Knowledge tracing (KT) defines the task of predicting whether students can correctly answer questions based on their historical response. Although much research has been devoted to exploiting the question information, plentiful advanced information among questions and skills hasn't been well extracted, making it challenging for previous work to perform adequately. In this paper, we demonstrate that large gains on KT can be realized by pre-training embeddings for each question on abundant side information, followed by training deep KT models on the obtained embeddings. To be specific, the side information includes question difficulty and three kinds of relations contained in a bipartite graph between questions and skills. To pre-train the question embeddings, we propose to use product-based neural networks to recover the side information. As a result, adopting the pre-trained embeddings in existing deep KT models significantly outperforms state-of-the-art baselines on three common KT datasets.

Yang Yang, Jian Shen, Yanru Qu et al.

With the rapid development in online education, knowledge tracing (KT) has become a fundamental problem which traces students' knowledge status and predicts their performance on new questions. Questions are often numerous in online education systems, and are always associated with much fewer skills. However, the previous literature fails to involve question information together with high-order question-skill correlations, which is mostly limited by data sparsity and multi-skill problems. From the model perspective, previous models can hardly capture the long-term dependency of student exercise history, and cannot model the interactions between student-questions, and student-skills in a consistent way. In this paper, we propose a Graph-based Interaction model for Knowledge Tracing (GIKT) to tackle the above probems. More specifically, GIKT utilizes graph convolutional network (GCN) to substantially incorporate question-skill correlations via embedding propagation. Besides, considering that relevant questions are usually scattered throughout the exercise history, and that question and skill are just different instantiations of knowledge, GIKT generalizes the degree of students' master of the question to the interactions between the student's current state, the student's history related exercises, the target question, and related skills. Experiments on three datasets demonstrate that GIKT achieves the new state-of-the-art performance, with at least 1% absolute AUC improvement.

Yunfei Liu, Xingjun Ma, James Bailey et al.

Recent studies have shown that DNNs can be compromised by backdoor attacks crafted at training time. A backdoor attack installs a backdoor into the victim model by injecting a backdoor pattern into a small proportion of the training data. At test time, the victim model behaves normally on clean test data, yet consistently predicts a specific (likely incorrect) target class whenever the backdoor pattern is present in a test example. While existing backdoor attacks are effective, they are not stealthy. The modifications made on training data or labels are often suspicious and can be easily detected by simple data filtering or human inspection. In this paper, we present a new type of backdoor attack inspired by an important natural phenomenon: reflection. Using mathematical modeling of physical reflection models, we propose reflection backdoor (Refool) to plant reflections as backdoor into a victim model. We demonstrate on 3 computer vision tasks and 5 datasets that, Refool can attack state-of-the-art DNNs with high success rate, and is resistant to state-of-the-art backdoor defenses.

Yunfei Liu, Yu Li, Shaodi You et al.

Intrinsic image decomposition, which is an essential task in computer vision, aims to infer the reflectance and shading of the scene. It is challenging since it needs to separate one image into two components. To tackle this, conventional methods introduce various priors to constrain the solution, yet with limited performance. Meanwhile, the problem is typically solved by supervised learning methods, which is actually not an ideal solution since obtaining ground truth reflectance and shading for massive general natural scenes is challenging and even impossible. In this paper, we propose a novel unsupervised intrinsic image decomposition framework, which relies on neither labeled training data nor hand-crafted priors. Instead, it directly learns the latent feature of reflectance and shading from unsupervised and uncorrelated data. To enable this, we explore the independence between reflectance and shading, the domain invariant content constraint and the physical constraint. Extensive experiments on both synthetic and real image datasets demonstrate consistently superior performance of the proposed method.

Yunfei Liu, Yu Li, Shaodi You et al.

Reflection is common in images capturing scenes behind a glass window, which is not only a disturbance visually but also influence the performance of other computer vision algorithms. Single image reflection removal is an ill-posed problem because the color at each pixel needs to be separated into two values, i.e., the desired clear background and the reflection. To solve it, existing methods propose priors such as smoothness, color consistency. However, the low-level priors are not reliable in complex scenes, for instance, when capturing a real outdoor scene through a window, both the foreground and background contain both smooth and sharp area and a variety of color. In this paper, inspired by the fact that human can separate the two layers easily by recognizing the objects, we use the object semantic as guidance to force the same semantic object belong to the same layer. Extensive experiments on different datasets show that adding the semantic information offers a significant improvement to reflection separation. We also demonstrate the applications of the proposed method to other computer vision tasks.

Yunfei Liu, Feng Lu

Many real world vision tasks, such as reflection removal from a transparent surface and intrinsic image decomposition, can be modeled as single image layer separation. However, this problem is highly ill-posed, requiring accurately aligned and hard to collect triplet data to train the CNN models. To address this problem, this paper proposes an unsupervised method that requires no ground truth data triplet in training. At the core of the method are two assumptions about data distributions in the latent spaces of different layers, based on which a novel unsupervised layer separation pipeline can be derived. Then the method can be constructed based on the GANs framework with self-supervision and cycle consistency constraints, etc. Experimental results demonstrate its successfulness in outperforming existing unsupervised methods in both synthetic and real world tasks. The method also shows its ability to solve a more challenging multi-layer separation task.

Huangyue Yu, Minjie Cai, Yunfei Liu et al.

In recent years, more and more videos are captured from the first-person viewpoint by wearable cameras. Such first-person video provides additional information besides the traditional third-person video, and thus has a wide range of applications. However, techniques for analyzing the first-person video can be fundamentally different from those for the third-person video, and it is even more difficult to explore the shared information from both viewpoints. In this paper, we propose a novel method for first- and third-person video co-analysis. At the core of our method is the notion of "joint attention", indicating the learnable representation that corresponds to the shared attention regions in different viewpoints and thus links the two viewpoints. To this end, we develop a multi-branch deep network with a triplet loss to extract the joint attention from the first- and third-person videos via self-supervised learning. We evaluate our method on the public dataset with cross-viewpoint video matching tasks. Our method outperforms the state-of-the-art both qualitatively and quantitatively. We also demonstrate how the learned joint attention can benefit various applications through a set of additional experiments.