arXiv subjects

Weijian Cao

Publications and source records attributed to Weijian Cao.

15 recordsLinked to original sources

arXiv2026

PersonaShot: Benchmarking Person-Centric Narrative Continuity in Multi-Shot Video Generation

Video generation is rapidly evolving from single-shot clips to multi-shot narratives, where the human character serves as the core narrative anchor. However, existing benchmarks mainly assess character appearance or individual-shot quality, without measuring whether physical and emotional states remain coherent across cuts. They also rarely provide criterion-specific evaluation methods, although physical continuity, facial dynamics, and cinematic relations require different visual, temporal, and relational evidence. To address these limitations, we introduce PersonaShot, the first person-centric benchmark for narrative continuity in multi-shot video generation. PersonaShot contains approximately 1,000 multi-shot segments and 16 metrics spanning physical continuity, affective dynamics, and cinematic grammar. \textbf{\textit{1)} Narrative Continuity Benchmark:} We evaluate character coherence across three temporal levels: within-shot states, cross-shot transitions, and sequence-level trajectories. \textbf{\textit{2)} Human-Aligned Specialist Evaluators:} We distill reasoning from a large multimodal teacher into lightweight criterion-specific evaluators, each grounded in the visual, temporal, or relational evidence required by its metric, and align them with expert human judgments. \textbf{\textit{3)} Systematic Evaluation and Insights:} Our evaluation reveals distinct capability profiles across state-of-the-art models and a clear gap between perceptual quality and cross-shot narrative continuity. Even visually compelling videos frequently exhibit physical-state resets, abrupt affective shifts, and broken cinematic relations across shots. Human studies further demonstrate strong agreement between our evaluators and expert judgments.

cs.CV↗

arXiv2026

VicEdit: Learning to Edit Videos from Visual In-Context Examples

Despite progress in instruction-based video editing, unimodal textual instructions inherently struggle to convey fine-grained textures and complex dynamics. To bridge this perceptual gap, we propose Visual In-context Editing, a new paradigm elevating video editing from textual instructions to multi-modal visual guidance encompassing single image, image pair, and video pair. To facilitate this paradigm, we curate VicEdit-400K, the first large-scale dataset for visual in-context video editing. We develop an automated pipeline to generate 400K high-quality samples across ten task types, ensuring superior visual fidelity and semantic consistency through multi-dimensional filtering. Leveraging this foundation, we introduce VicEdit, a unified framework to bridge visual and textual contexts. To adaptively extract editing semantics from heterogeneous references, we design Modality-Adaptive Semantic Distillation, which produces modality-specific semantic tokens from visual references. These tokens are then synergistically integrated with textual instructions through Dual-Context Injection, enabling the generation process to benefit from both visual and textual signals. Extensive evaluations on VicEditBench demonstrate that VicEdit achieves state-of-the-art performance across both basic instruction editing and visual in-context editing tasks, establishing visual in-context learning as a powerful and controllable paradigm for video editing.

cs.CV↗

arXiv2026

Semantic Frame Interpolation

Generating intermediate video content of varying lengths based on given first and last frames, along with text prompt information, offers significant research and application potential. However, traditional frame interpolation tasks primarily focus on scenarios with a small number of frames, no text control, and minimal differences between the first and last frames. Recent community developers have utilized large video models represented by Wan to endow frame-to-frame capabilities. However, these models can only generate a fixed number of frames and often fail to produce satisfactory results for certain frame lengths, while this setting lacks a clear official definition and a well-established benchmark. In this paper, we first propose a new practical Semantic Frame Interpolation (SFI) task from the perspective of academic definition, which covers the above two settings and supports inference at multiple frame rates. To achieve this goal, we propose a novel SemFi model building upon Wan2.1, which incorporates a Mixture-of-LoRA module to ensure the generation of high-consistency content that aligns with control conditions across various frame length limitations. Furthermore, we propose SFI-300K, the first general-purpose dataset and benchmark specifically designed for SFI. To support this, we collect and process data from the perspective of SFI, carefully designing evaluation metrics and methods to assess the model's performance across multiple dimensions, encompassing image and video, and various aspects, including consistency and diversity. Through extensive experiments on SFI-300K, we demonstrate that our method is particularly well-suited to meet the requirements of the SFI task.

cs.CV↗

arXiv2025

Transform Trained Transformer: Accelerating Naive 4K Video Generation Over 10$\times$

Native 4K (2160$\times$3840) video generation remains a critical challenge due to the quadratic computational explosion of full-attention as spatiotemporal resolution increases, making it difficult for models to strike a balance between efficiency and quality. This paper proposes a novel Transformer retrofit strategy termed $\textbf{T3}$ ($\textbf{T}$ransform $\textbf{T}$rained $\textbf{T}$ransformer) that, without altering the core architecture of full-attention pretrained models, significantly reduces compute requirements by optimizing their forward logic. Specifically, $\textbf{T3-Video}$ introduces a multi-scale weight-sharing window attention mechanism and, via hierarchical blocking together with an axis-preserving full-attention design, can effect an "attention pattern" transformation of a pretrained model using only modest compute and data. Results on 4K-VBench show that $\textbf{T3-Video}$ substantially outperforms existing approaches: while delivering performance improvements (+4.29$\uparrow$ VQA and +0.08$\uparrow$ VTC), it accelerates native 4K video generation by more than 10$\times$. Project page at https://zhangzjn.github.io/projects/T3-Video

cs.CV↗

arXiv2025

Soul: Breathe Life into Digital Human for High-fidelity Long-term Multimodal Animation

We propose a multimodal-driven framework for high-fidelity long-term digital human animation termed $\textbf{Soul}$, which generates semantically coherent videos from a single-frame portrait image, text prompts, and audio, achieving precise lip synchronization, vivid facial expressions, and robust identity preservation. We construct Soul-1M, containing 1 million finely annotated samples with a precise automated annotation pipeline (covering portrait, upper-body, full-body, and multi-person scenes) to mitigate data scarcity, and we carefully curate Soul-Bench for comprehensive and fair evaluation of audio-/text-guided animation methods. The model is built on the Wan2.2-5B backbone, integrating audio-injection layers and multiple training strategies together with threshold-aware codebook replacement to ensure long-term generation consistency. Meanwhile, step/CFG distillation and a lightweight VAE are used to optimize inference efficiency, achieving an 11.4$\times$ speedup with negligible quality loss. Extensive experiments show that Soul significantly outperforms current leading open-source and commercial models on video quality, video-text alignment, identity preservation, and lip-synchronization accuracy, demonstrating broad applicability in real-world scenarios such as virtual anchors and film production. Project page at https://zhangzjn.github.io/projects/Soul/

cs.CV↗

arXiv2025

Collaborative Face Experts Fusion in Video Generation: Boosting Identity Consistency Across Large Face Poses

Current video generation models struggle with identity preservation under large face poses, primarily facing two challenges: the difficulty in exploring an effective mechanism to integrate identity features into DiT architectures, and the lack of targeted coverage of large face poses in existing open-source video datasets. To address these, we present two key innovations. First, we propose Collaborative Face Experts Fusion (CoFE), which dynamically fuses complementary signals from three specialized experts within the DiT backbone: an identity expert that captures cross-pose invariant features, a semantic expert that encodes high-level visual context, and a detail expert that preserves pixel-level attributes such as skin texture and color gradients. Second, we introduce a data curation pipeline comprising three key components: Face Constraints to ensure diverse large-pose coverage, Identity Consistency to maintain stable identity across frames, and Speech Disambiguation to align textual captions with actual speaking behavior. This pipeline yields LaFID-180K, a large-scale dataset of pose-annotated video clips designed for identity-preserving video generation. Experimental results on several benchmarks demonstrate that our approach significantly outperforms state-of-the-art methods in face similarity, FID, and CLIP semantic alignment. Project page: https://rain152.github.io/CoFE/.

cs.CV↗

arXiv2025

Identity-Preserving Text-to-Video Generation Guided by Simple yet Effective Spatial-Temporal Decoupled Representations

Identity-preserving text-to-video (IPT2V) generation, which aims to create high-fidelity videos with consistent human identity, has become crucial for downstream applications. However, current end-to-end frameworks suffer a critical spatial-temporal trade-off: optimizing for spatially coherent layouts of key elements (e.g., character identity preservation) often compromises instruction-compliant temporal smoothness, while prioritizing dynamic realism risks disrupting the spatial coherence of visual structures. To tackle this issue, we propose a simple yet effective spatial-temporal decoupled framework that decomposes representations into spatial features for layouts and temporal features for motion dynamics. Specifically, our paper proposes a semantic prompt optimization mechanism and stage-wise decoupled generation paradigm. The former module decouples the prompt into spatial and temporal components. Aligned with the subsequent stage-wise decoupled approach, the spatial prompts guide the text-to-image (T2I) stage to generate coherent spatial features, while the temporal prompts direct the sequential image-to-video (I2V) stage to ensure motion consistency. Experimental results validate that our approach achieves excellent spatiotemporal consistency, demonstrating outstanding performance in identity preservation, text relevance, and video quality. By leveraging this simple yet robust mechanism, our algorithm secures the runner-up position in 2025 ACM MultiMedia Challenge. Our code is available at https://github.com/rain152/IPVG.

cs.CV↗

arXiv2025

SwiftVideo: A Unified Framework for Few-Step Video Generation through Trajectory-Distribution Alignment

Diffusion-based or flow-based models have achieved significant progress in video synthesis but require multiple iterative sampling steps, which incurs substantial computational overhead. While many distillation methods that are solely based on trajectory-preserving or distribution-matching have been developed to accelerate video generation models, these approaches often suffer from performance breakdown or increased artifacts under few-step settings. To address these limitations, we propose \textbf{\emph{SwiftVideo}}, a unified and stable distillation framework that combines the advantages of trajectory-preserving and distribution-matching strategies. Our approach introduces continuous-time consistency distillation to ensure precise preservation of ODE trajectories. Subsequently, we propose a dual-perspective alignment that includes distribution alignment between synthetic and real data along with trajectory alignment across different inference steps. Our method maintains high-quality video generation while substantially reducing the number of inference steps. Quantitative evaluations on the OpenVid-1M benchmark demonstrate that our method significantly outperforms existing approaches in few-step video generation.

cs.CV↗

arXiv2025

StrandDesigner: Towards Practical Strand Generation with Sketch Guidance

Realistic hair strand generation is crucial for applications like computer graphics and virtual reality. While diffusion models can generate hairstyles from text or images, these inputs lack precision and user-friendliness. Instead, we propose the first sketch-based strand generation model, which offers finer control while remaining user-friendly. Our framework tackles key challenges, such as modeling complex strand interactions and diverse sketch patterns, through two main innovations: a learnable strand upsampling strategy that encodes 3D strands into multi-scale latent spaces, and a multi-scale adaptive conditioning mechanism using a transformer with diffusion heads to ensure consistency across granularity levels. Experiments on several benchmark datasets show our method outperforms existing approaches in realism and precision. Qualitative results further confirm its effectiveness. Code will be released at [GitHub](https://github.com/fighting-Zhang/StrandDesigner).

cs.CV↗

arXiv2024

ID-Sculpt: ID-aware 3D Head Generation from Single In-the-wild Portrait Image

While recent works have achieved great success on image-to-3D object generation, high quality and fidelity 3D head generation from a single image remains a great challenge. Previous text-based methods for generating 3D heads were limited by text descriptions and image-based methods struggled to produce high-quality head geometry. To handle this challenging problem, we propose a novel framework, ID-Sculpt, to generate high-quality 3D heads while preserving their identities. Our work incorporates the identity information of the portrait image into three parts: 1) geometry initialization, 2) geometry sculpting, and 3) texture generation stages. Given a reference portrait image, we first align the identity features with text features to realize ID-aware guidance enhancement, which contains the control signals representing the face information. We then use the canny map, ID features of the portrait image, and a pre-trained text-to-normal/depth diffusion model to generate ID-aware geometry supervision, and 3D-GAN inversion is employed to generate ID-aware geometry initialization. Furthermore, with the ability to inject identity information into 3D head generation, we use ID-aware guidance to calculate ID-aware Score Distillation (ISD) for geometry sculpting. For texture generation, we adopt the ID Consistent Texture Inpainting and Refinement which progressively expands the view for texture inpainting to obtain an initialization UV texture map. We then use the ID-aware guidance to provide image-level supervision for noisy multi-view images to obtain a refined texture map. Extensive experiments demonstrate that we can generate high-quality 3D heads with accurate geometry and texture from a single in-the-wild portrait image.

cs.CV↗

arXiv2024

MambaGesture: Enhancing Co-Speech Gesture Generation with Mamba and Disentangled Multi-Modality Fusion

Co-speech gesture generation is crucial for producing synchronized and realistic human gestures that accompany speech, enhancing the animation of lifelike avatars in virtual environments. While diffusion models have shown impressive capabilities, current approaches often overlook a wide range of modalities and their interactions, resulting in less dynamic and contextually varied gestures. To address these challenges, we present MambaGesture, a novel framework integrating a Mamba-based attention block, MambaAttn, with a multi-modality feature fusion module, SEAD. The MambaAttn block combines the sequential data processing strengths of the Mamba model with the contextual richness of attention mechanisms, enhancing the temporal coherence of generated gestures. SEAD adeptly fuses audio, text, style, and emotion modalities, employing disentanglement to deepen the fusion process and yield gestures with greater realism and diversity. Our approach, rigorously evaluated on the multi-modal BEAT dataset, demonstrates significant improvements in Fréchet Gesture Distance (FGD), diversity scores, and beat alignment, achieving state-of-the-art performance in co-speech gesture generation. Project website: $\href{https://fcchit.github.io/mambagesture/}{\textit{https://fcchit.github.io/mambagesture/}}$.

cs.HC↗

arXiv2024

VividPose: Advancing Stable Video Diffusion for Realistic Human Image Animation

Human image animation involves generating a video from a static image by following a specified pose sequence. Current approaches typically adopt a multi-stage pipeline that separately learns appearance and motion, which often leads to appearance degradation and temporal inconsistencies. To address these issues, we propose VividPose, an innovative end-to-end pipeline based on Stable Video Diffusion (SVD) that ensures superior temporal stability. To enhance the retention of human identity, we propose an identity-aware appearance controller that integrates additional facial information without compromising other appearance details such as clothing texture and background. This approach ensures that the generated videos maintain high fidelity to the identity of human subject, preserving key facial features across various poses. To accommodate diverse human body shapes and hand movements, we introduce a geometry-aware pose controller that utilizes both dense rendering maps from SMPL-X and sparse skeleton maps. This enables accurate alignment of pose and shape in the generated videos, providing a robust framework capable of handling a wide range of body shapes and dynamic hand movements. Extensive qualitative and quantitative experiments on the UBCFashion and TikTok benchmarks demonstrate that our method achieves state-of-the-art performance. Furthermore, VividPose exhibits superior generalization capabilities on our proposed in-the-wild dataset. Codes and models will be available.

cs.CV↗

arXiv2024

FreeMotion: A Unified Framework for Number-free Text-to-Motion Synthesis

Text-to-motion synthesis is a crucial task in computer vision. Existing methods are limited in their universality, as they are tailored for single-person or two-person scenarios and can not be applied to generate motions for more individuals. To achieve the number-free motion synthesis, this paper reconsiders motion generation and proposes to unify the single and multi-person motion by the conditional motion distribution. Furthermore, a generation module and an interaction module are designed for our FreeMotion framework to decouple the process of conditional motion generation and finally support the number-free motion synthesis. Besides, based on our framework, the current single-person motion spatial control method could be seamlessly integrated, achieving precise control of multi-person motion. Extensive experiments demonstrate the superior performance of our method and our capability to infer single and multi-human motions simultaneously.

cs.CV↗

arXiv2024

DiffFAE: Advancing High-fidelity One-shot Facial Appearance Editing with Space-sensitive Customization and Semantic Preservation

Facial Appearance Editing (FAE) aims to modify physical attributes, such as pose, expression and lighting, of human facial images while preserving attributes like identity and background, showing great importance in photograph. In spite of the great progress in this area, current researches generally meet three challenges: low generation fidelity, poor attribute preservation, and inefficient inference. To overcome above challenges, this paper presents DiffFAE, a one-stage and highly-efficient diffusion-based framework tailored for high-fidelity FAE. For high-fidelity query attributes transfer, we adopt Space-sensitive Physical Customization (SPC), which ensures the fidelity and generalization ability by utilizing rendering texture derived from 3D Morphable Model (3DMM). In order to preserve source attributes, we introduce the Region-responsive Semantic Composition (RSC). This module is guided to learn decoupled source-regarding features, thereby better preserving the identity and alleviating artifacts from non-facial attributes such as hair, clothes, and background. We further introduce a consistency regularization for our pipeline to enhance editing controllability by leveraging prior knowledge in the attention matrices of diffusion model. Extensive experiments demonstrate the superiority of DiffFAE over existing methods, achieving state-of-the-art performance in facial appearance editing.

cs.CV↗

arXiv2024

TexDreamer: Towards Zero-Shot High-Fidelity 3D Human Texture Generation

Texturing 3D humans with semantic UV maps remains a challenge due to the difficulty of acquiring reasonably unfolded UV. Despite recent text-to-3D advancements in supervising multi-view renderings using large text-to-image (T2I) models, issues persist with generation speed, text consistency, and texture quality, resulting in data scarcity among existing datasets. We present TexDreamer, the first zero-shot multimodal high-fidelity 3D human texture generation model. Utilizing an efficient texture adaptation finetuning strategy, we adapt large T2I model to a semantic UV structure while preserving its original generalization capability. Leveraging a novel feature translator module, the trained model is capable of generating high-fidelity 3D human textures from either text or image within seconds. Furthermore, we introduce ArTicuLated humAn textureS (ATLAS), the largest high-resolution (1024 X 1024) 3D human texture dataset which contains 50k high-fidelity textures with text descriptions.

cs.CV↗