Existing feedforward subject-driven video customization methods mainly study single-subject scenarios due to the difficulty of constructing multi-subject training data pairs. Another challenging problem that how to use the signals such as depth, mask, camera, and text prompts to control and edit the subject in the customized video is still less explored. In this paper, we first propose a data construction pipeline, VideoCus-Factory, to produce training data pairs for multi-subject customization from raw videos without labels and control signals such as depth-to-video and mask-to-video pairs. Based on our constructed data, we develop an Image-Video Transfer Mixed (IVTM) training with image editing data to enable instructive editing for the subject in the customized video. Then we propose a diffusion Transformer framework, OmniVCus, with two embedding mechanisms, Lottery Embedding (LE) and Temporally Aligned Embedding (TAE). LE enables inference with more subjects by using the training subjects to activate more frame embeddings. TAE encourages the generation process to extract guidance from temporally aligned control signals by assigning the same frame embeddings to the control and noise tokens. Experiments demonstrate that our method significantly surpasses state-of-the-art methods in both quantitative and qualitative evaluations. Project page is at https://caiyuanhao1998.github.io/project/OmniVCus/

Despite the remarkable generation quality of video Diffusion Transformer (DiT) models, their practical deployment is severely hindered by extensive computational requirements. This inefficiency stems from two key challenges: the quadratic complexity of self-attention with respect to token length and the multi-step nature of diffusion models. To address these limitations, we present Jenga, a novel inference pipeline that combines dynamic attention carving with progressive resolution generation. Our approach leverages two key insights: (1) early denoising steps do not require high-resolution latents, and (2) later steps do not require dense attention. Jenga introduces a block-wise attention mechanism that dynamically selects relevant token interactions using 3D space-filling curves, alongside a progressive resolution strategy that gradually increases latent resolution during generation. Experimental results demonstrate that Jenga achieves substantial speedups across multiple state-of-the-art video diffusion models while maintaining comparable generation quality (8.83$\times$ speedup with 0.01\% performance drop on VBench). As a plug-and-play solution, Jenga enables practical, high-quality video generation on modern hardware by reducing inference time from minutes to seconds---without requiring model retraining.

Text-based video editing has recently attracted considerable interest in changing the style or replacing the objects with a similar structure. Beyond this, we demonstrate that properties such as shape, size, location, motion, etc., can also be edited in videos. Our key insight is that the keyframe's transformations of the specific internal feature (e.g., edge maps of objects or human pose), can easily propagate to other frames to provide generation guidance. We thus propose MagicStick, a controllable video editing method that edits the video properties by utilizing the transformation on the extracted internal control signals. In detail, to keep the appearance, we inflate both the pre-trained image diffusion model and ControlNet to the temporal dimension and train low-rank adaptions (LoRA) layers to fit the specific scenes. Then, in editing, we perform an inversion and editing framework. Differently, finetuned ControlNet is introduced in both inversion and generation for attention guidance with the proposed attention remix between the spatial attention maps of inversion and editing. Yet succinct, our method is the first method to show the ability of video property editing from the pre-trained text-to-image model. We present experiments on numerous examples within our unified framework. We also compare with shape-aware text-based editing and handcrafted motion video generation, demonstrating our superior temporal consistency and editing capability than previous works.

We present a framework to generate past and future processes for drawing process videos. Given a canvas image uploaded by a user, the framework can generate both preceding and succeeding states of the drawing process, and the generated states can be reused as inputs for further state generation. We observe that the user queries typically have one-to-one or many-to-many states, and in many cases, involve non-contiguous states. This necessitates a backend that solves a set-to-set problem with arbitrary combinations of past or future states. To this end, we repurpose video diffusion models to learn the set-to-set mapping with pretrained video priors. We implement the system with strong diffusion transformer backbones ( e.g. , CogVideoX and LTXVideo) and high-quality data processing ( e.g. , sampling short shots from long videos of real drawing records). Experiments show that the generated states are diverse in drawing contexts and resemble human drawing processes. This capability may aid artists in visualizing potential outcomes, generating creative inspirations, or refining existing workflows.

Creating a vivid video from the event or scenario in our imagination is a truly fascinating experience. Recent advancements in text-to-video synthesis have unveiled the potential to achieve this with prompts only. While text is convenient in conveying the overall scene context, it may be insufficient to control precisely. In this paper, we explore customized video generation by utilizing text as context description and motion structure (e.g., frame-wise depth) as concrete guidance. Our method, dubbed Make-Your-Video, involves joint-conditional video generation using a Latent Diffusion Model that is pre-trained for still image synthesis and then promoted for video generation with the introduction of temporal modules. This two-stage learning scheme not only reduces the computing resources required, but also improves the performance by transferring the rich concepts available in image datasets solely into video generation. Moreover, we use a simple yet effective causal attention mask strategy to enable longer video synthesis, which mitigates the potential quality degradation effectively. Experimental results show the superiority of our method over existing baselines, particularly in terms of temporal coherence and fidelity to users' guidance. In addition, our model enables several intriguing applications that demonstrate potential for practical usage.

Text-to-video (T2V) models have shown remarkable capabilities in generating diverse videos. However, they struggle to produce user-desired artistic videos due to (i) text's inherent clumsiness in expressing specific styles and (ii) the generally degraded style fidelity. To address these challenges, we introduce StyleCrafter, a generic method that enhances pretrained T2V models with a style control adapter, allowing video generation in any style by feeding a reference image. Considering the scarcity of artistic video data, we propose to first train a style control adapter using style-rich image datasets, then transfer the learned stylization ability to video generation through a tailor-made finetuning paradigm. To promote content-style disentanglement, we employ carefully designed data augmentation strategies to enhance decoupled learning. Additionally, we propose a scale-adaptive fusion module to balance the influences of text-based content features and image-based style features, which helps generalization across various text and style combinations. StyleCrafter efficiently generates high-quality stylized videos that align with the content of the texts and resemble the style of the reference images. Experiments demonstrate that our approach is more flexible and efficient than existing competitors. Project page: https://gongyeliu.github.io/StyleCrafter.github.io/

We introduce ToonCrafter, a novel approach that transcends traditional correspondence-based cartoon video interpolation, paving the way for generative interpolation. Traditional methods, that implicitly assume linear motion and the absence of complicated phenomena like dis-occlusion, often struggle with the exaggerated non-linear and large motions with occlusion commonly found in cartoons, resulting in implausible or even failed interpolation results. To overcome these limitations, we explore the potential of adapting live-action video priors to better suit cartoon interpolation within a generative framework. ToonCrafter effectively addresses the challenges faced when applying live-action video motion priors to generative cartoon interpolation. First, we design a toon rectification learning strategy that seamlessly adapts live-action video priors to the cartoon domain, resolving the domain gap and content leakage issues. Next, we introduce a dual-reference-based 3D decoder to compensate for lost details due to the highly compressed latent prior spaces, ensuring the preservation of fine details in interpolation results. Finally, we design a flexible sketch encoder that empowers users with interactive control over the interpolation results. Experimental results demonstrate that our proposed method not only produces visually convincing and more natural dynamics, but also effectively handles dis-occlusion. The comparative evaluation demonstrates the notable superiority of our approach over existing competitors. Code and model weights are available at https://doubiiu.github.io/projects/ToonCrafter

Speech-driven 3D facial animation has been widely studied, yet there is still a gap to achieving realism and vividness due to the highly ill-posed nature and scarcity of audio-visual data. Existing works typically formulate the cross-modal mapping into a regression task, which suffers from the regression-to-mean problem leading to over-smoothed facial motions. In this paper, we propose to cast speech-driven facial animation as a code query task in a finite proxy space of the learned codebook, which effectively promotes the vividness of the generated motions by reducing the cross-modal mapping uncertainty. The codebook is learned by self-reconstruction over real facial motions and thus embedded with realistic facial motion priors. Over the discrete motion space, a temporal autoregressive model is employed to sequentially synthesize facial motions from the input speech signal, which guarantees lip-sync as well as plausible facial expressions. We demonstrate that our approach outperforms current state-of-the-art methods both qualitatively and quantitatively. Also, a user study further justifies our superiority in perceptual quality. Code and video demo are available at https://doubiiu.github.io/projects/codetalker.

Current 3D scene stylization methods transfer textures and colors as styles using arbitrary style references, lacking meaningful semantic correspondences. We introduce Reference-Based Non-Photorealistic Radiance Fields (Ref-NP R) to address this limitation. This controllable method stylizes a 3D scene using radiance fields with a single stylized 2D view as a reference. We propose a ray registration process based on the stylized reference view to obtain pseudo-ray supervision in novel views. Then we exploit semantic correspondences in content images to fill occluded regions with perceptually similar styles, resulting in non-photorealistic and continuous novel view sequences. Our experimental results demonstrate that Ref-NPR out-performs existing scene and video stylization methods regarding visual quality and semantic correspondence. The code and data are publicly available on the project page at https://ref-npr.github.io.

Recent diffusion model advancements have enabled high-fidelity images to be generated using text prompts. However, a domain gap exists between generated images and real-world images, which poses a challenge in generating high-quality variations of real-world images. Our investigation uncovers that this domain gap originates from a latents' distribution gap in different diffusion processes. To address this issue, we propose a novel inference pipeline called Real-world Image Variation by ALignment (RIVAL) that utilizes diffusion models to generate image variations from a single image exemplar. Our pipeline enhances the generation quality of image variations by aligning the image generation process to the source image's inversion chain. Specifically, we demonstrate that step-wise latent distribution alignment is essential for generating high-quality variations. To attain this, we design a cross-image self-attention injection for feature interaction and a step-wise distribution normalization to align the latent features. Incorporating these alignment processes into a diffusion model allows RIVAL to generate high-quality image variations without further parameter optimization. Our experimental results demonstrate that our proposed approach outperforms existing methods concerning semantic similarity and perceptual quality. This generalized inference pipeline can be easily applied to other diffusion-based generation tasks, such as image-conditioned text-to-image generation and stylization. Project page: https://rival-diff.github.io

Conventional social media platforms usually downscale high-resolution (HR) images to restrict their resolution to a specific size for saving transmission/storage cost, which makes those visual details inaccessible to other users. To bypass this obstacle, recent invertible image downscaling methods jointly model the downscaling/upscaling problems and achieve impressive performance. However, they only consider fixed integer scale factors and may be inapplicable to generic downscaling tasks towards resolution restriction as posed by social media platforms. In this paper, we propose an effective and universal Scale-Arbitrary Invertible Image Downscaling Network (AIDN), to downscale HR images with arbitrary scale factors in an invertible manner. Particularly, the HR information is embedded in the downscaled low-resolution (LR) counterparts in a nearly imperceptible form such that our AIDN can further restore the original HR images solely from the LR images. The key to supporting arbitrary scale factors is our proposed Conditional Resampling Module (CRM) that conditions the downscaling/upscaling kernels and sampling locations on both scale factors and image content. Extensive experimental results demonstrate that our AIDN achieves top performance for invertible downscaling with both arbitrary integer and non-integer scale factors. Also, both quantitative and qualitative evaluations show our AIDN is robust to the lossy image compression standard. The source code and trained models are publicly available at https://github.com/Doubiiu/AIDN.

A popular and challenging task in video research, frame interpolation aims to increase the frame rate of video. Most existing methods employ a fixed motion model, e.g., linear, quadratic, or cubic, to estimate the intermediate warping field. However, such fixed motion models cannot well represent the complicated non-linear motions in the real world or rendered animations. Instead, we present an adaptive flow prediction module to better approximate the complex motions in video. Furthermore, interpolating just one intermediate frame between consecutive input frames may be insufficient for complicated non-linear motions. To enable multi-frame interpolation, we introduce the time as a control variable when interpolating frames between original ones in our generic adaptive flow prediction module. Qualitative and quantitative experimental results show that our method can produce high-quality results and outperforms the existing state-of-the-art methods on popular public datasets.