Transformer-based methods have shown impressive performance in image restoration tasks, such as image super-resolution and denoising. However, we find that these networks can only utilize a limited spatial range of input information through attribution analysis. This implies that the potential of Transformer is still not fully exploited in existing networks. In order to activate more input pixels for better restoration, we propose a new Hybrid Attention Transformer (HAT). It combines both channel attention and window-based self-attention schemes, thus making use of their complementary advantages. Moreover, to better aggregate the cross-window information, we introduce an overlapping cross-attention module to enhance the interaction between neighboring window features. In the training stage, we additionally adopt a same-task pre-training strategy to further exploit the potential of the model for further improvement. Extensive experiments have demonstrated the effectiveness of the proposed modules. We further scale up the model to show that the performance of the SR task can be greatly improved. Besides, we extend HAT to more image restoration applications, including real-world image super-resolution, Gaussian image denoising and image compression artifacts reduction. Experiments on benchmark and real-world datasets demonstrate that our HAT achieves state-of-the-art performance both quantitatively and qualitatively.

Filmmaking and animation production often require sophisticated techniques for coordinating camera transitions and object movements, typically involving labor-intensive real-world capturing. Despite advancements in generative AI for video creation, achieving precise control over motion for interactive video asset generation remains challenging. To this end, we propose Image Conductor, a method for precise control of camera transitions and object movements to generate video assets from a single image. An well-cultivated training strategy is proposed to separate distinct camera and object motion by camera LoRA weights and object LoRA weights. To further eliminate motion ambiguity from ill-posed trajectories, we introduce a camera-free guidance technique during inference process, enhancing object movements while eliminating camera transitions. Additionally, we develop a trajectory-oriented video motion data curation pipeline for training. Quantitative and qualitative experiments demonstrate our method's precision and fine-grained control in generating motion-controllable videos from images, advancing the practical application of interactive video synthesis.

Although diffusion-based techniques have shown remarkable success in image generation and editing tasks, their abuse can lead to severe negative social impacts. Recently, some works have been proposed to provide defense against the abuse of diffusion-based methods. However, their protection may be limited in specific scenarios by manually defined prompts or the stable diffusion (SD) version. Furthermore, these methods solely focus on tuning methods, overlooking editing methods that could also pose a significant threat. In this work, we propose Anti-Diffusion, a privacy protection system designed for general diffusion-based methods, applicable to both tuning and editing techniques. To mitigate the limitations of manually defined prompts on defense performance, we introduce the prompt tuning (PT) strategy that enables precise expression of original images. To provide defense against both tuning and editing methods, we propose the semantic disturbance loss (SDL) to disrupt the semantic information of protected images. Given the limited research on the defense against editing methods, we develop a dataset named Defense-Edit to assess the defense performance of various methods. Experiments demonstrate that our Anti-Diffusion achieves superior defense performance across a wide range of diffusion-based techniques in different scenarios.

Customized video generation aims to generate high-quality videos guided by text prompts and subject's reference images. However, since it is only trained on static images, the fine-tuning process of subject learning disrupts abilities of video diffusion models (VDMs) to combine concepts and generate motions. To restore these abilities, some methods use additional video similar to the prompt to fine-tune or guide the model. This requires frequent changes of guiding videos and even re-tuning of the model when generating different motions, which is very inconvenient for users. In this paper, we propose CustomCrafter, a novel framework that preserves the model's motion generation and conceptual combination abilities without additional video and fine-tuning to recovery. For preserving conceptual combination ability, we design a plug-and-play module to update few parameters in VDMs, enhancing the model's ability to capture the appearance details and the ability of concept combinations for new subjects. For motion generation, we observed that VDMs tend to restore the motion of video in the early stage of denoising, while focusing on the recovery of subject details in the later stage. Therefore, we propose Dynamic Weighted Video Sampling Strategy. Using the pluggability of our subject learning modules, we reduce the impact of this module on motion generation in the early stage of denoising, preserving the ability to generate motion of VDMs. In the later stage of denoising, we restore this module to repair the appearance details of the specified subject, thereby ensuring the fidelity of the subject's appearance. Experimental results show that our method has a significant improvement compared to previous methods.

Pre-trained video generation models hold great potential for generative video super-resolution (VSR). However, adapting them for full-size VSR, as most existing methods do, suffers from unnecessary intensive full-attention computation and fixed output resolution. To overcome these limitations, we make the first exploration into utilizing video diffusion priors for patch-wise VSR. This is non-trivial because pre-trained video diffusion models are not native for patch-level detail generation. To mitigate this challenge, we propose an innovative approach, called PatchVSR, which integrates a dual-stream adapter for conditional guidance. The patch branch extracts features from input patches to maintain content fidelity while the global branch extracts context features from the resized full video to bridge the generation gap caused by incomplete semantics of patches. Particularly, we also inject the patch’s location information into the model to better contextualize patch synthesis within the global video frame. Experiments demonstrate that our method can synthesize high-fidelity, high-resolution details at the patch level. A tailor-made multi-patch joint modulation is proposed to ensure visual consistency across individually enhanced patches. Due to the flexibility of our patch-based paradigm, we can achieve highly competitive 4K VSR based on a 512×512 resolution base model, with extremely high efficiency.

Camera control has been actively studied in text or image conditioned video generation tasks. However, altering camera trajectories of a given video remains under-explored, despite its importance in the field of video creation. It is non-trivial due to the extra constraints of maintaining multiple-frame appearance and dynamic synchronization. To address this, we present ReCamMaster, a camera-controlled generative video re-rendering framework that reproduces the dynamic scene of an input video at novel camera trajectories. The core innovation lies in harnessing the generative capabilities of pre-trained text-to-video models through a simple yet powerful video conditioning mechanism--its capability is often overlooked in current research. To overcome the scarcity of qualified training data, we construct a comprehensive multi-camera synchronized video dataset using Unreal Engine 5, which is carefully curated to follow real-world filming characteristics, covering diverse scenes and camera movements. It helps the model generalize to in-the-wild videos. Lastly, we further improve the robustness to diverse inputs through a meticulously designed training strategy. Extensive experiments show that our method substantially outperforms existing state-of-the-art approaches. Our method also finds promising applications in video stabilization, super-resolution, and outpainting. Our code and dataset are publicly available at: https://github.com/KwaiVGI/ReCamMaster.

Recent advancements in video diffusion models demonstrate remarkable capabilities in simulating real-world dynamics and 3D consistency. This progress motivates us to explore the potential of these models to maintain dynamic consistency across diverse viewpoints, a feature highly sought after in applications like virtual filming. Unlike existing methods focused on multi-view generation of single objects for 4D reconstruction, our interest lies in generating open-world videos from arbitrary viewpoints, incorporating six degrees of freedom (6 DoF) camera poses. To achieve this, we propose a plug-and-play module that enhances a pre-trained text-to-video model for multi-camera video generation, ensuring consistent content across different viewpoints. Specifically, we introduce a multi-view synchronization module designed to maintain appearance and geometry consistency across these viewpoints. Given the scarcity of high-quality training data, we also propose a progressive training scheme that leverages multi-camera images and monocular videos as a supplement to Unreal Engine-rendered multi-camera videos. This comprehensive approach significantly benefits our model. Experimental results demonstrate the superiority of our proposed method over existing competitors and several baselines. Furthermore, our method enables intriguing extensions, such as re-rendering a video from multiple novel viewpoints. Project webpage: https://jianhongbai.github.io/SynCamMaster/

This paper aims to manipulate multi-entity 3D motions in video generation. Previous methods on controllable video generation primarily leverage 2D control signals to manipulate object motions and have achieved remarkable synthesis results. However, 2D control signals are inherently limited in expressing the 3D nature of object motions. To overcome this problem, we introduce 3DTrajMaster, a robust controller that regulates multi-entity dynamics in 3D space, given user-desired 6DoF pose (location and rotation) sequences of entities. At the core of our approach is a plug-and-play 3D-motion grounded object injector that fuses multiple input entities with their respective 3D trajectories through a gated self-attention mechanism. In addition, we exploit an injector architecture to preserve the video diffusion prior, which is crucial for generalization ability. To mitigate video quality degradation, we introduce a domain adaptor during training and employ an annealed sampling strategy during inference. To address the lack of suitable training data, we construct a 360-Motion Dataset, which first correlates collected 3D human and animal assets with GPT-generated trajectory and then captures their motion with 12 evenly-surround cameras on diverse 3D UE platforms. Extensive experiments show that 3DTrajMaster sets a new state-of-the-art in both accuracy and generalization for controlling multi-entity 3D motions. Project page: http://fuxiao0719.github.io/projects/3dtrajmaster

We propose Flow-GRPO, the first method to integrate online policy gradient reinforcement learning (RL) into flow matching models. Our approach uses two key strategies: (1) an ODE-to-SDE conversion that transforms a deterministic Ordinary Differential Equation (ODE) into an equivalent Stochastic Differential Equation (SDE) that matches the original model's marginal distribution at all timesteps, enabling statistical sampling for RL exploration; and (2) a Denoising Reduction strategy that reduces training denoising steps while retaining the original number of inference steps, significantly improving sampling efficiency without sacrificing performance. Empirically, Flow-GRPO is effective across multiple text-to-image tasks. For compositional generation, RL-tuned SD3.5-M generates nearly perfect object counts, spatial relations, and fine-grained attributes, increasing GenEval accuracy from $63\%$ to $95\%$. In visual text rendering, accuracy improves from $59\%$ to $92\%$, greatly enhancing text generation. Flow-GRPO also achieves substantial gains in human preference alignment. Notably, very little reward hacking occurred, meaning rewards did not increase at the cost of appreciable image quality or diversity degradation.

Video generation has achieved significant advances through rectified flow techniques, but issues like unsmooth motion and misalignment between videos and prompts persist. In this work, we develop a systematic pipeline that harnesses human feedback to mitigate these problems and refine the video generation model. Specifically, we begin by constructing a large-scale human preference dataset focused on modern video generation models, incorporating pairwise annotations across multi-dimensions. We then introduce VideoReward, a multi-dimensional video reward model, and examine how annotations and various design choices impact its rewarding efficacy. From a unified reinforcement learning perspective aimed at maximizing reward with KL regularization, we introduce three alignment algorithms for flow-based models. These include two training-time strategies: direct preference optimization for flow (Flow-DPO) and reward weighted regression for flow (Flow-RWR), and an inference-time technique, Flow-NRG, which applies reward guidance directly to noisy videos. Experimental results indicate that VideoReward significantly outperforms existing reward models, and Flow-DPO demonstrates superior performance compared to both Flow-RWR and supervised fine-tuning methods. Additionally, Flow-NRG lets users assign custom weights to multiple objectives during inference, meeting personalized video quality needs.

The advent of Multimodal Large Language Models (MLLMs) has expanded AI capabilities to visual modalities, yet existing evaluation benchmarks remain limited to single-video understanding, overlooking the critical need for multi-video understanding in real-world scenarios (e.g., sports analytics and autonomous driving). To address this significant gap, we introduce MVU-Eval, the first comprehensive benchmark for evaluating Multi-Video Understanding for MLLMs. Specifically, our MVU-Eval mainly assesses eight core competencies through 1,824 meticulously curated question-answer pairs spanning 4,959 videos from diverse domains, addressing both fundamental perception tasks and high-order reasoning tasks. These capabilities are rigorously aligned with real-world applications such as multi-sensor synthesis in autonomous systems and cross-angle sports analytics. Through extensive evaluation of state-of-the-art open-source and closed-source models, we reveal significant performance discrepancies and limitations in current MLLMs' ability to perform understanding across multiple videos.The benchmark will be made publicly available to foster future research.

Camera control is crucial for generating expressive and cinematic videos. Existing methods rely on explicit sequences of camera parameters as control conditions, which can be cumbersome for users to construct, particularly for intricate camera movements. To provide a more intuitive camera control method, we propose CamCloneMaster, a framework that enables users to replicate camera movements from reference videos without requiring camera parameters or test-time fine-tuning. CamCloneMaster seamlessly supports reference-based camera control for both Image-to-Video and Video-to-Video tasks within a unified framework. Furthermore, we present the Camera Clone Dataset, a large-scale synthetic dataset designed for camera clone learning, encompassing diverse scenes, subjects, and camera movements. Extensive experiments and user studies demonstrate that CamCloneMaster outperforms existing methods in terms of both camera controllability and visual quality. Dataset and Code can be found at https://camclonemaster.github.io/.

Recent advances in interactive video generation have shown promising results, yet existing approaches struggle with scene-consistent memory capabilities in long video generation due to limited use of historical context. In this work, we propose Context-as-Memory, which utilizes historical context as memory for video generation. It includes two simple yet effective designs: (1) storing context in frame format without additional post-processing; (2) conditioning by concatenating context and frames to be predicted along the frame dimension at the input, requiring no external control modules. Furthermore, considering the enormous computational overhead of incorporating all historical context, we propose the Memory Retrieval module to select truly relevant context frames by determining FOV (Field of View) overlap between camera poses, which significantly reduces the number of candidate frames without substantial information loss. Experiments demonstrate that Context-as-Memory achieves superior memory capabilities in interactive long video generation compared to SOTAs, even generalizing effectively to open-domain scenarios not seen during training. Our project page are publicly available at https://context-as-memory.github.io/.

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.

Generating text-editable and pose-controllable character videos have an imperious demand in creating various digital human. Nevertheless, this task has been restricted by the absence of a comprehensive dataset featuring paired video-pose captions and the generative prior models for videos. In this work, we design a novel two-stage training scheme that can utilize easily obtained datasets (i.e., image pose pair and pose-free video) and the pre-trained text-to-image (T2I) model to obtain the pose-controllable character videos. Specifically, in the first stage, only the keypoint image pairs are used only for a controllable text-to-image generation. We learn a zero-initialized convolutional encoder to encode the pose information. In the second stage, we finetune the motion of the above network via a pose-free video dataset by adding the learnable temporal self-attention and reformed cross-frame self-attention blocks. Powered by our new designs, our method successfully generates continuously pose-controllable character videos while keeps the editing and concept composition ability of the pre-trained T2I model. The code and models are available on https://follow-your-pose.github.io/.

The incredible generative ability of large-scale text-to-image (T2I) models has demonstrated strong power of learning complex structures and meaningful semantics. However, relying solely on text prompts cannot fully take advantage of the knowledge learned by the model, especially when flexible and accurate controlling (e.g., structure and color) is needed. In this paper, we aim to ``dig out" the capabilities that T2I models have implicitly learned, and then explicitly use them to control the generation more granularly. Specifically, we propose to learn low-cost T2I-Adapters to align internal knowledge in T2I models with external control signals, while freezing the original large T2I models. In this way, we can train various adapters according to different conditions, achieving rich control and editing effects in the color and structure of the generation results. Further, the proposed T2I-Adapters have attractive properties of practical value, such as composability and generalization ability. Extensive experiments demonstrate that our T2I-Adapter has promising generation quality and a wide range of applications. Our code is available at https://github.com/TencentARC/T2I-Adapter.

Controllable spherical panoramic image generation holds substantial applicative potential across a variety of domains. However, it remains a challenging task due to the inherent spherical distortion and geometry characteristics, resulting in low-quality content generation. In this paper, we introduce a novel framework of SphereDiffusion to address these unique challenges, for better generating high-quality and precisely controllable spherical panoramic images. For the spherical distortion characteristic, we embed the semantics of the distorted object with text encoding, then explicitly construct the relationship with text-object correspondence to better use the pre-trained knowledge of the planar images. Meanwhile, we employ a deformable technique to mitigate the semantic deviation in latent space caused by spherical distortion. For the spherical geometry characteristic, in virtue of spherical rotation invariance, we improve the data diversity and optimization objectives in the training process, enabling the model to better learn the spherical geometry characteristic. Furthermore, we enhance the denoising process of the diffusion model, enabling it to effectively use the learned geometric characteristic to ensure the boundary continuity of the generated images. With these specific techniques, experiments on Structured3D dataset show that SphereDiffusion significantly improves the quality of controllable spherical image generation and relatively reduces around 35% FID on average.

Text-to-video generation aims to produce a video based on a given prompt. Recently, several commercial video models have been able to generate plausible videos with mini-mal noise, excellent details, and high aesthetic scores. However, these models rely on large-scale, well-filtered, high-quality videos that are not accessible to the community. Many existing research works, which train models using the low-quality WebVid-10M dataset, struggle to generate high-quality videos because the models are optimized to fit WebVid-10M. In this work, we explore the training scheme of video models extended from Stable Diffusion and investigate the feasibility of leveraging low-quality videos and synthesized high-quality images to obtain a high-quality video model. We first analyze the connection between the spatial and temporal modules of video models and the distribution shift to low-quality videos. We observe that full training of all modules results in a stronger coupling between spatial and temporal modules than only training temporal modules. Based on this stronger coupling, we shift the distribution to higher quality without motion degradation by finetuning spatial modules with high-quality images, resulting in a generic high-quality video model. Evaluations are conducted to demonstrate the superiority of the proposed method, particularly in picture quality, motion, and concept composition.

Vector-Quantized (VQ-based) generative models usually consist of two basic components, i.e., VQ tokenizers and generative transformers. Prior research focuses on improving the reconstruction fidelity of VQ tokenizers but rarely examines how the improvement in reconstruction affects the generation ability of generative transformers. In this paper, we surprisingly find that improving the reconstruction fidelity of VQ tokenizers does not necessarily improve the generation. Instead, learning to compress semantic features within VQ tokenizers significantly improves generative transformers' ability to capture textures and structures. We thus highlight two competing objectives of VQ tokeniz-ers for image synthesis: semantic compression and details preservation. Different from previous work that pri-oritizes better details preservation, we propose Semantic-Quantized GAN (SeQ-GAN) with two learning phases to balance the two objectives. In the first phase, we propose a semantic-enhanced perceptual loss for better semantic compression. In the second phase, we fix the encoder and codebook, but enhance and finetune the decoder to achieve better details preservation. Our proposed SeQ-GAN significantly improves VQ-based generative models for both un-conditional and conditional image generation. Specifically, SeQ-GAN achieves a Fré chet Inception Distance (FID) of 6.25 and Inception Score (IS) of 140.9 on 256×256 Ima-geNet generation, which is a remarkable improvement over VIT-VQGAN (714M), which obtains 11.2 FID and 97.2 IS.

Current instruction-based image editing methods, such as InstructPix2Pix, often fail to produce satisfactory results in complex scenarios due to their dependence on the simple CLIP text encoder in diffusion models. To rectify this, this paper introduces SmartEdit, a novel approach of instruction-based image editing that leverages Multimodal Large Language Models (MLLMs) to enhance its understanding and reasoning capabilities. However, direct integration of these elements still faces challenges in situations requiring complex reasoning. To mitigate this, we propose a Bidirectional Interaction Module (BIM) that enables comprehensive bidirectional information interactions between the input image and the MLLM output. During training, we initially incorporate perception data to boost the perception and understanding capabilities of diffusion models. Subsequently, we demonstrate that a small amount of complex instruction editing data can effectively stimulate SmartEdit’ s editing capabilities for more complex instructions. We further construct a new evaluation dataset, Reason-Edit, specifically tailored for complex instruction-based image editing. Both quantitative and qualitative results on this evaluation dataset indicate that our SmartEdit surpasses previous methods, paving the way for the practical application of complex instruction-based image editing.

Recent advances in text-to-image generation have made remarkable progress in synthesizing realistic human photos conditioned on given text prompts. However, existing per-sonalized generation methods cannot simultaneously sat-isfy the requirements of high efficiency, promising identity (ID) fidelity, and flexible text controllability. In this work, we introduce PhotoMaker, an efficient personalized text-to-image generation method, which mainly encodes an arbitrary number of input ID images into a stack ID embed-ding for preserving ID information. Such an embedding, serving as a unified ID representation, can not only encap-sulate the characteristics of the same input ID comprehen-sively, but also accommodate the characteristics of differ-ent IDs for subsequent integration. This paves the way for more intriguing and practically valuable applications. Be-sides, to drive the training of our PhotoMaker, we propose an ID-oriented data construction pipeline to assemble the training data. Under the nourishment of the dataset constructed through the proposed pipeline, our PhotoMaker demonstrates better ID preservation ability than test-time fine-tuning based methods, yet provides significant speed improvements, high-quality generation results, strong gen-eralization capabilities, and a wide range of applications.

The vision and language generative models have been overgrown in recent years. For video generation, various open-sourced models and public-available services have been developed to generate high-quality videos. However, these methods often use a few metrics, e.g., FVD [56] or IS [45], to evaluate the performance. We argue that it is hard to judge the large conditional generative models from the simple metrics since these models are often trained on very large datasets with multi-aspect abilities. Thus, we propose a novel framework and pipeline for exhaustively evaluating the performance of the generated videos. Our approach involves generating a diverse and comprehensive list of 700 prompts for text-to-video generation, which is based on an analysis of real-world user data and generated with the assistance of a large language model. Then, we evaluate the state-of-the-art video generative models on our carefully designed benchmark, in terms of visual qualities, content qualities, motion qualities, and text-video alignment with 17 well-selected objective metrics. To obtain the finalleaderboard of the models, we further fit a series of coefficients to align the objective metrics to the users' opinions. Based on the proposed human alignment method, our final score shows a higher correlation than simply averaging the metrics, showing the effectiveness of the proposed evaluation method.

Large-scale Text-to-Image (T2I) diffusion models have revolutionized image generation over the last few years. Although owning diverse and high-quality generation capabilities, translating these abilities to fine-grained Image editing remains challenging. In this paper, we propose DiffEditor to rectify two weaknesses in existing diffusion-based image editing: (1) in complex scenarios, editing results often lack editing accuracy and exhibit unexpected artifacts; (2) lack of flexibility to harmonize editing operations, e.g., imagine new content. In our solution, we introduce image prompts in fine-grained image editing, cooperating with the text prompt to better describe the editing content. To increase the flexibility while maintaining content consistency, we locally combine stochastic differential equation (SDE) into the ordinary differential equation (ODE) sampling. In addition, we incorporate regional score-based gradient guidance and a time travel strategy into the diffusion sampling, further improving the editing quality. Extensive experiments demonstrate that our method can efficiently achieve state-of-the-art performance on various fine-grained image editing tasks, including editing within a single image (e.g., object moving, resizing, and content dragging) and across images (e.g., appearance replacing and object pasting). Our source code is released at https://github.com/MC-E/DragonDiffusion.

We introduce X-Adapter, a universal upgrader to enable the pretrained plug-and-play modules (e.g., ControlNet, LoRA) to work directly with the upgraded text-to-image diffusion model (e.g., SDXL) without further retraining. We achieve this goal by training an additional network to control the frozen upgraded model with the new text-image data pairs. In detail, X-Adapter keeps a frozen copy of the old model to preserve the connectors of different plugins. Additionally, X-Adapter adds trainable mapping layers that bridge the decoders from models of different versions for feature remapping. The remapped features will be used as guidance for the upgraded model. To enhance the guidance ability of X-Adapter, we employ a null-text training strategy for the upgraded model. After training, we also introduce a two-stage denoising strategy to align the initial latents of X Adapter and the upgraded model. Thanks to our strategies, X-Adapter demonstrates universal compatibility with various plugins and also enables plugins of different versions to work together, thereby expanding the functionalities of diffusion community. To verify the effectiveness of the proposed method, we conduct extensive experiments and the results show that X-Adapter may facilitate wider application in the upgraded foundational diffusion model. Project page at: https://showlab.github.io/X-Adapter/.

Video and audio content creation serves as the core technique for the movie industry and professional users. Re-cently, existing diffusion-based methods tackle video and audio generation separately, which hinders the technique transfer from academia to industry. In this work, we aim at filling the gap, with a carefully designed optimization-based framework for cross-visual-audio and joint-visual-audio generation. We observe the powerful generation abil-ity of off-the-shelf video or audio generation models. Thus, instead of training the giant models from scratch, we pro-pose to bridge the existing strong models with a shared la-tent representation space. Specifically, we propose a mul-timodality latent aligner with the pre-trained ImageBind model. Our latent aligner shares a similar core as the clas-sifier guidance that guides the diffusion denoising process during inference time. Through carefully designed opti-mization strategy and loss functions, we show the superior performance of our method on joint video-audio generation, visual-steered audio generation, and audio-steered vi-sual generation tasks. The project website can be found at https://yzxing87.github.io/Seeing-and-Hearing/.

We introduce SUPIR (Scaling-UP Image Restoration), a groundbreaking image restoration method that harnesses generative prior and the power of model scaling up. Lever-aging multi-modal techniques and advanced generative prior, SUPIR marks a significant advance in intelligent and realistic image restoration. As a pivotal catalyst within SUPIR, model scaling dramatically enhances its capabil-ities and demonstrates new potential for image restoration. We collect a dataset comprising 20 million high-resolution, high-quality images for model training, each en-riched with descriptive text annotations. SUPIR provides the capability to restore images guided by textual prompts, broadening its application scope and potential. Moreover, we introduce negative-quality prompts to further improve perceptual quality. We also develop a restoration-guided sampling method to suppress the fidelity issue encountered in generative-based restoration. Experiments demonstrate SUPIR's exceptional restoration effects and its novel capac-ity to manipulate restoration through textual prompts.

The great success of Large Language Models (LLMs) has expanded the potential of multimodality, contributing to the gradual evolution of General Artificial Intelligence (AGI). A true AGI agent should not only possess the capability to perform predefined multi-tasks but also exhibit emergent abilities in an open-world context. However, despite the considerable advancements made by recent multimodal LLMs, they still fall short in effectively unifying comprehension and generation tasks, let alone open-world emergent abilities. We contend that the key to overcoming the present impasse lies in enabling text and images to be represented and processed interchangeably within a unified autoregressive Transformer. To this end, we introduce $\textbf{SEED}$, an elaborate image tokenizer that empowers LLMs with the ability to $\textbf{SEE}$ and $\textbf{D}$raw at the same time. We identify two crucial design principles: (1) Image tokens should be independent of 2D physical patch positions and instead be produced with a $\textit{1D causal dependency}$, exhibiting intrinsic interdependence that aligns with the left-to-right autoregressive prediction mechanism in LLMs. (2) Image tokens should capture $\textit{high-level semantics}$ consistent with the degree of semantic abstraction in words, and be optimized for both discriminativeness and reconstruction during the tokenizer training phase. With SEED tokens, LLM is able to perform scalable multimodal autoregression under its original training recipe, i.e., next-word prediction. SEED-LLaMA is therefore produced by large-scale pretraining and instruction tuning on the interleaved textual and visual data, demonstrating impressive performance on a broad range of multimodal comprehension and generation tasks. More importantly, SEED-LLaMA has exhibited compositional emergent abilities such as multi-turn in-context multimodal generation, acting like your AI assistant. The code (training and inference) and models are released in https://github.com/AILab-CVC/SEED.

In this work, we investigate the capability of generating images from pre-trained diffusion models at much higher resolutions than the training image sizes. In addition, the generated images should have arbitrary image aspect ratios. When generating images directly at a higher resolution, 1024 x 1024, with the pre-trained Stable Diffusion using training images of resolution 512 x 512, we observe persistent problems of object repetition and unreasonable object structures. Existing works for higher-resolution generation, such as attention-based and joint-diffusion approaches, cannot well address these issues. As a new perspective, we examine the structural components of the U-Net in diffusion models and identify the crucial cause as the limited perception field of convolutional kernels. Based on this key observation, we propose a simple yet effective re-dilation that can dynamically adjust the convolutional perception field during inference. We further propose the dispersed convolution and noise-damped classifier-free guidance, which can enable ultra-high-resolution image generation (e.g., 4096 x 4096). Notably, our approach does not require any training or optimization. Extensive experiments demonstrate that our approach can address the repetition issue well and achieve state-of-the-art performance on higher-resolution image synthesis, especially in texture details. Our work also suggests that a pre-trained diffusion model trained on low-resolution images can be directly used for high-resolution visual generation without further tuning, which may provide insights for future research on ultra-high-resolution image and video synthesis. More results are available at the anonymous website: https://scalecrafter.github.io/ScaleCrafter/

Despite the ability of text-to-image (T2I) diffusion models to generate high-quality images, transferring this ability to accurate image editing remains a challenge. In this paper, we propose a novel image editing method, DragonDiffusion, enabling Drag-style manipulation on Diffusion models. Specifically, we treat image editing as the change of feature correspondence in a pre-trained diffusion model. By leveraging feature correspondence, we develop energy functions that align with the editing target, transforming image editing operations into gradient guidance. Based on this guidance approach, we also construct multi-scale guidance that considers both semantic and geometric alignment. Furthermore, we incorporate a visual cross-attention strategy based on a memory bank design to ensure consistency between the edited result and original image. Benefiting from these efficient designs, all content editing and consistency operations come from the feature correspondence without extra model fine-tuning. Extensive experiments demonstrate that our method has promising performance on various image editing tasks, including within a single image (e.g., object moving, resizing, and content dragging) or across images (e.g., appearance replacing and object pasting). Code is available at https://github.com/MC-E/DragonDiffusion.

With the availability of large-scale video datasets and the advances of diffusion models, text-driven video generation has achieved substantial progress. However, existing video generation models are typically trained on a limited number of frames, resulting in the inability to generate high-fidelity long videos during inference. Furthermore, these models only support single-text conditions, whereas real-life scenarios often require multi-text conditions as the video content changes over time. To tackle these challenges, this study explores the potential of extending the text-driven capability to generate longer videos conditioned on multiple texts. 1) We first analyze the impact of initial noise in video diffusion models. Then building upon the observation of noise, we propose FreeNoise, a tuning-free and time-efficient paradigm to enhance the generative capabilities of pretrained video diffusion models while preserving content consistency. Specifically, instead of initializing noises for all frames, we reschedule a sequence of noises for long-range correlation and perform temporal attention over them by window-based fusion. 2) Additionally, we design a novel motion injection method to support the generation of videos conditioned on multiple text prompts. Extensive experiments validate the superiority of our paradigm in extending the generative capabilities of video diffusion models. It is noteworthy that compared with the previous best-performing method which brought about 255% extra time cost, our method incurs only negligible time cost of approximately 17%. Generated video samples are available at our website: http://haonanqiu.com/projects/FreeNoise.html.

Image processing is a fundamental task in computer vision, which aims at enhancing image quality and extracting essential features for subsequent vision applications. Traditionally, task-specific models are developed for individual tasks and designing such models requires distinct expertise. Building upon the success of large language models (LLMs) in natural language processing (NLP), there is a similar trend in computer vision, which focuses on developing large-scale models through pretraining and in-context learning. This paradigm shift reduces the reliance on task-specific models, yielding a powerful unified model to deal with various tasks. However, these advances have predominantly concentrated on high-level vision tasks, with less attention paid to low-level vision tasks. To address this issue, we propose a universal model for general image processing that covers image restoration, image enhancement, image feature extraction tasks, etc. Our proposed framework, named PromptGIP, unifies these diverse image processing tasks within a universal framework. Inspired by NLP question answering (QA) techniques, we employ a visual prompting question answering paradigm. Specifically, we treat the input-output image pair as a structured question-answer sentence, thereby reprogramming the image processing task as a prompting QA problem. PromptGIP can undertake diverse cross-domain tasks using provided visual prompts, eliminating the need for task-specific finetuning. Capable of handling up to 15 different image processing tasks, PromptGIP represents a versatile and adaptive approach to general image processing. While PromptGIP has demonstrated a certain degree of out-of-domain task generalization capability, further research is expected to fully explore its more powerful emergent generalization. Codes will be available at https://github.com/lyh-18/PromptGIP.

Incorporating a customized object into image generation presents an attractive feature in text-to-image (T2I) generation. Some methods finetune T2I models for each object individually at test-time, which tend to be overfitted and time-consuming. Others train an extra encoder to extract object visual information for customization efficiently but struggle to preserve the object's identity. To address these limitations, we present CustomNet, a unified encoder-based object customization framework that explicitly incorporates 3D novel view synthesis capabilities into the customization process. This integration facilitates the adjustment of spatial positions and viewpoints, producing diverse outputs while effectively preserving the object's identity. To train our model effectively, we propose a dataset construction pipeline to better handle real-world objects and complex backgrounds. Additionally, we introduce delicate designs that enable location control and flexible background control through textual descriptions or user-defined backgrounds. Our method allows for object customization without the need of test-time optimization, providing simultaneous control over viewpoints, location, and text. Experimental results show that our method outperforms other customization methods regarding identity preservation, diversity, and harmony. Codes are available at https://github.com/TencentARC/CustomNet.

Sora's high-motion intensity and long consistent videos have significantly impacted the field of video generation, attracting unprecedented attention. However, existing publicly available datasets are inadequate for generating Sora-like videos, as they mainly contain short videos with low motion intensity and brief captions. To address these issues, we propose MiraData, a high-quality video dataset that surpasses previous ones in video duration, caption detail, motion strength, and visual quality. We curate MiraData from diverse, manually selected sources and meticulously process the data to obtain semantically consistent clips. GPT-4V is employed to annotate structured captions, providing detailed descriptions from four different perspectives along with a summarized dense caption. To better assess temporal consistency and motion intensity in video generation, we introduce MiraBench, which enhances existing benchmarks by adding 3D consistency and tracking-based motion strength metrics. MiraBench includes 150 evaluation prompts and 17 metrics covering temporal consistency, motion strength, 3D consistency, visual quality, text-video alignment, and distribution similarity. To demonstrate the utility and effectiveness of MiraData, we conduct experiments using our DiT-based video generation model, MiraDiT. The experimental results on MiraBench demonstrate the superiority of MiraData, especially in motion strength.

Despite significant advancements in video generation and editing using diffusion models, achieving accurate and localized video editing remains a substantial challenge. Additionally, most existing video editing methods primarily focus on altering visual content, with limited research dedicated to motion editing. In this paper, we present a novel attempt to Remake a Video (ReVideo) which stands out from existing methods by allowing precise video editing in specific areas through the specification of both content and motion. Content editing is facilitated by modifying the first frame, while the trajectory-based motion control offers an intuitive user interaction experience. ReVideo addresses a new task involving the coupling and training imbalance between content and motion control. To tackle this, we develop a three-stage training strategy that progressively decouples these two aspects from coarse to fine. Furthermore, we propose a spatiotemporal adaptive fusion module to integrate content and motion control across various sampling steps and spatial locations. Extensive experiments demonstrate that our ReVideo has promising performance on several accurate video editing applications, i.e., (1) locally changing video content while keeping the motion constant, (2) keeping content unchanged and customizing new motion trajectories, (3) modifying both content and motion trajectories. Our method can also seamlessly extend these applications to multi-area editing without specific training, demonstrating its flexibility and robustness.

Diffusion models have demonstrated great success in text-to-video (T2V) generation. However, existing methods may face challenges when handling complex (long) video generation scenarios that involve multiple objects or dynamic changes in object numbers. To address these limitations, we propose VideoTetris, a novel framework that enables compositional T2V generation. Specifically, we propose spatio-temporal compositional diffusion to precisely follow complex textual semantics by manipulating and composing the attention maps of denoising networks spatially and temporally. Moreover, we propose a new dynamic-aware data processing pipeline and a consistency regularization method to enhance the consistency of auto-regressive video generation. Extensive experiments demonstrate that our VideoTetris achieves impressive qualitative and quantitative results in compositional T2V generation. Code is available at: https://github.com/YangLing0818/VideoTetris

Video colorization is a challenging and highly ill-posed problem. Although recent years have witnessed remarkable progress in single image colorization, there is relatively less research effort on video colorization, and existing methods always suffer from severe flickering artifacts (temporal inconsistency) or unsatisfactory colorization. We address this problem from a new perspective, by jointly considering colorization and temporal consistency in a unified framework. Specifically, we propose a novel temporally consistent video colorization (TCVC) framework. TCVC effectively propagates frame-level deep features in a bidirectional way to enhance the temporal consistency of colorization. Furthermore, TCVC introduces a self-regularization learning (SRL) scheme to minimize the differences in predictions obtained using different time steps. SRL does not require any ground-truth color videos for training and can further improve temporal consistency. Experiments demonstrate that our method can not only provide visually pleasing colorized video, but also with clearly better temporal consistency than state-of-the-art methods. A video demo is provided at https://www.youtube.com/watch?v=c7dczMs-olE , while code is available at https://github.com/lyh-18/TCVC-Temporally-Consistent-Video-Colorization .

Interactive image restoration aims to construct an interactive pathway between users and restoration networks, which empowers users to modulate the restoration results according to their own demands. However, existing methods are primarily limited to training their networks with predefined and simplistic synthetic degradations. Consequently, these methods often encounter significant performance degradation when confronted with real-world degradations that deviate from their assumptions. Furthermore, existing interactive image restoration approaches solely support global modulation, wherein a single modulation factor governs the reconstruction process for the entire image. In this paper, we propose a novel method to perform real-world and intricate image super-resolution in an interactive manner. Specifically, we propose a metric-learning-based degradation estimation strategy to estimate not only the overall degradation level of the entire image but also the finer-grained, pixel-wise degradation within real-world scenarios. This enables local control over the restoration results by selectively modulating the corresponding regions based on the densely-estimated degradation map. Additionally, a new metric-argumented loss is proposed to further enhance the performance of real-world image super-resolution. Through extensive experimentation, we demonstrate the efficacy of our method in achieving exceptional modulation and restoration performance in real-world image super-resolution tasks, all while maintaining an appealing model complexity.

Recent progress in blind face restoration has resulted in producing high-quality restored results for static images. However, efforts to extend these advancements to video scenarios have been minimal, partly because of the absence of benchmarks that allow for a comprehensive and fair comparison. In this work, we first present a fair evaluation benchmark, in which we first introduce a Real-world Low-Quality Face Video benchmark (RFV-LQ), evaluate several leading image-based face restoration algorithms, and conduct a thorough systematical analysis of the benefits and challenges associated with extending blind face image restoration algorithms to degraded face videos. Our analysis identifies several key issues, primarily categorized into two aspects: significant jitters in facial components and noise-shape flickering between frames. To address these issues, we propose a Temporal Consistency Network (TCN) cooperated with alignment smoothing to reduce jitters and flickers in restored videos. TCN is a flexible component that can be seamlessly plugged into the most advanced face image restoration algorithms, ensuring the quality of image-based restoration is maintained as closely as possible. Extensive experiments have been conducted to evaluate the effectiveness and efficiency of our proposed TCN and alignment smoothing operation.

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

Despite that convolution neural networks (CNN) have recently demonstrated high-quality reconstruction for video super-resolution (VSR), efficiently training competitive VSR models remains a challenging problem. It usually takes an order of magnitude more time than training their counterpart image models, leading to long research cycles. Existing VSR methods typically train models with fixed spatial and temporal sizes from beginning to end. The fixed sizes are usually set to large values for good performance, resulting to slow training. However, is such a rigid training strategy necessary for VSR? In this work, we show that it is possible to gradually train video models from small to large spatial/temporal sizes, \ie, in an easy-to-hard manner. In particular, the whole training is divided into several stages and the earlier stage has smaller training spatial shape. Inside each stage, the temporal size also varies from short to long while the spatial size remains unchanged. Training is accelerated by such a multigrid training strategy, as most of computation is performed on smaller spatial and shorter temporal shapes. For further acceleration with GPU parallelization, we also investigate the large minibatch training without the loss in accuracy. Extensive experiments demonstrate that our method is capable of largely speeding up training (up to $6.2\times$ speedup in wall-clock training time) without performance drop for various VSR models.

The recurrent structure is a prevalent framework for the task of video super-resolution, which models the temporal dependency between frames via hidden states. When applied to real-world scenarios with unknown and complex degradations, hidden states tend to contain unpleasant artifacts and propagate them to restored frames. In this circumstance, our analyses show that such artifacts can be largely alleviated when the hidden state is replaced with a cleaner counterpart. Based on the observations, we propose a Hidden State Attention (HSA) module to mitigate artifacts in real-world video super-resolution. Specifically, we first adopt various cheap filters to produce a hidden state pool. For example, Gaussian blur filters are for smoothing artifacts while sharpening filters are for enhancing details. To aggregate a new hidden state that contains fewer artifacts from the hidden state pool, we devise a Selective Cross Attention (SCA) module, in which the attention between input features and each hidden state is calculated. Equipped with HSA, our proposed method, namely FastRealVSR, is able to achieve 2x speedup while obtaining better performance than Real-BasicVSR. Codes will be available at https://github.com/TencentARC/FastRealVSR.

Transformer-based methods have shown impressive performance in low-level vision tasks, such as image super-resolution. However, we find that these networks can only utilize a limited spatial range of input information through attribution analysis. This implies that the potential of Transformer is still not fully exploited in existing networks. In order to activate more input pixels for better reconstruction, we propose a novel Hybrid Attention Transformer (HAT). It combines both channel attention and window-based self-attention schemes, thus making use of their complementary advantages of being able to utilize global statistics and strong local fitting capability. Moreover, to better aggregate the cross-window information, we introduce an overlapping cross-attention module to enhance the interaction between neighboring window features. In the training stage, we additionally adopt a same-task pre-training strategy to exploit the potential of the model for further improvement. Extensive experiments show the effectiveness of the proposed modules, and we further scale up the model to demonstrate that the performance of this task can be greatly improved. Our overall method significantly outperforms the state-of-the-art methods by more than 1dB.

Recent CLIP-guided 3D optimization methods, such as DreamFields [19] and PureCLIPNeRF [24], have achieved impressive results in zero-shot text-to-3D synthesis. However, due to scratch training and random initialization without prior knowledge, these methods often fail to generate accurate and faithful 3D structures that conform to the input text. In this paper, we make the first attempt to introduce explicit 3D shape priors into the CLIP-guided 3D optimization process. Specifically, we first generate a high-quality 3D shape from the input text in the text-to-shape stage as a 3D shape prior. We then use it as the initialization of a neural radiance field and optimize it with the full prompt. To address the challenging text-to-shape generation task, we present a simple yet effective approach that directly bridges the text and image modalities with a powerful text-to-image diffusion model. To narrow the style domain gap between the images synthesized by the text-to-image diffusion model and shape renderings used to train the image-to-shape generator, we further propose to jointly optimize a learnable text prompt and fine-tune the text-to-image diffusion model for rendering-style image generation. Our method, Dream3D, is capable of generating imaginative 3D content with superior visual quality and shape accuracy compared to state-of-the-art methods. Our project page is at https://bluestyle97.github.io/dream3d/.

Omnidirectional images (ODIs) have obtained lots of research interest for immersive experiences. Although ODIs require extremely high resolution to capture details of the entire scene, the resolutions of most ODIs are insufficient. Previous methods attempt to solve this issue by image super-resolution (SR) on equirectangular projection (ERP) images. However, they omit geometric properties of ERP in the degradation process, and their models can hardly generalize to real ERP images. In this paper, we propose Fisheye downsampling, which mimics the real-world imaging process and synthesizes more realistic low-resolution samples. Then we design a distortion-aware Transformer (OSRT) to modulate ERP distortions continuously and self-adaptively. Without a cumbersome process, OSRT outperforms previous methods by about 0.2dB on PSNR. Moreover, we propose a convenient data augmentation strategy, which synthesizes pseudo ERP images from plain images. This simple strategy can alleviate the over-fitting problem of large networks and significantly boost the performance of ODISR. Extensive experiments have demonstrated the state-of-the-art performance of our OSRT.

Despite the success in large-scale text-to-image generation and text-conditioned image editing, existing methods still struggle to produce consistent generation and editing results. For example, generation approaches usually fail to synthesize multiple images of the same objects/characters but with different views or poses. Meanwhile, existing editing methods either fail to achieve effective complex nonrigid editing while maintaining the overall textures and identity, or require time-consuming fine-tuning to capture the image-specific appearance. In this paper, we develop MasaCtrl, a tuning-free method to achieve consistent image generation and complex non-rigid image editing simultaneously. Specifically, MasaCtrl converts existing self-attention in diffusion models into mutual self-attention, so that it can query correlated local contents and textures from source images for consistency. To further alleviate the query confusion between foreground and background, we propose a mask-guided mutual self-attention strategy, where the mask can be easily extracted from the cross-attention maps. Extensive experiments show that the proposed MasaCtrl can produce impressive results in both consistent image generation and complex non-rigid real image editing.

The diffusion-based generative models have achieved remarkable success in text-based image generation. However, since it contains enormous randomness in generation progress, it is still challenging to apply such models for real-world visual content editing, especially in videos. In this paper, we propose FateZero, a zero-shot text-based editing method on real-world videos without per-prompt training or use-specific mask. To edit videos consistently, we propose several techniques based on the pre-trained models. Firstly, in contrast to the straightforward DDIM inversion technique, our approach captures intermediate attention maps during inversion, which effectively retain both structural and motion information. These maps are directly fused in the editing process rather than generated during denoising. To further minimize semantic leakage of the source video, we then fuse self-attentions with a blending mask obtained by cross-attention features from the source prompt. Furthermore, we have implemented a reform of the self-attention mechanism in denoising UNet by introducing spatial-temporal attention to ensure frame consistency. Yet succinct, our method is the first one to show the ability of zero-shot text-driven video style and local attribute editing from the trained text-to-image model. We also have a better zero-shot shape-aware editing ability based on the text-to-video model [52]. Extensive experiments demonstrate our superior temporal consistency and editing capability than previous works.

To replicate the success of text-to-image (T2I) generation, recent works employ large-scale video datasets to train a text-to-video (T2V) generator. Despite their promising results, such paradigm is computationally expensive. In this work, we propose a new T2V generation setting—One-Shot Video Tuning, where only one text-video pair is presented. Our model is built on state-of-the-art T2I diffusion models pre-trained on massive image data. We make two key observations: 1) T2I models can generate still images that represent verb terms; 2) extending T2I models to generate multiple images concurrently exhibits surprisingly good content consistency. To further learn continuous motion, we introduce Tune-A-Video, which involves a tailored spatio-temporal attention mechanism and an efficient one-shot tuning strategy. At inference, we employ DDIM inversion to provide structure guidance for sampling. Extensive qualitative and numerical experiments demonstrate the remarkable ability of our method across various applications.

Image super-resolution (SR) with generative adversarial networks (GAN) has achieved great success in restoring realistic details. However, it is notorious that GAN-based SR models will inevitably produce unpleasant and undesirable artifacts, especially in practical scenarios. Previous works typically suppress artifacts with an extra loss penalty in the training phase. They only work for in-distribution artifact types generated during training. When applied in real-world scenarios, we observe that those improved methods still generate obviously annoying artifacts during inference. In this paper, we analyze the cause and characteristics of the GAN artifacts produced in unseen test data without ground-truths. We then develop a novel method, namely, DeSRA, to Detect and then “Delete” those SR Artifacts in practice. Specifically, we propose to measure a relative local variance distance from MSE-SR results and GAN-SR results, and locate the problematic areas based on the above distance and semantic-aware thresholds. After detecting the artifact regions, we develop a finetune procedure to improve GAN-based SR models with a few samples, so that they can deal with similar types of artifacts in more unseen real data. Equipped with our DeSRA, we can successfully eliminate artifacts from inference and improve the ability of SR models to be applied in real-world scenarios. The code will be available at https://github.com/TencentARC/DeSRA.

Public large-scale text-to-image diffusion models, such as Stable Diffusion, have gained significant attention from the community. These models can be easily customized for new concepts using low-rank adaptations (LoRAs). However, the utilization of multiple-concept LoRAs to jointly support multiple customized concepts presents a challenge. We refer to this scenario as decentralized multi-concept customization, which involves single-client concept tuning and center-node concept fusion. In this paper, we propose a new framework called Mix-of-Show that addresses the challenges of decentralized multi-concept customization, including concept conflicts resulting from existing single-client LoRA tuning and identity loss during model fusion. Mix-of-Show adopts an embedding-decomposed LoRA (ED-LoRA) for single-client tuning and gradient fusion for the center node to preserve the in-domain essence of single concepts and support theoretically limitless concept fusion. Additionally, we introduce regionally controllable sampling, which extends spatially controllable sampling (e.g., ControlNet and T2I-Adapter) to address attribute binding and missing object problems in multi-concept sampling. Extensive experiments demonstrate that Mix-of-Show is capable of composing multiple customized concepts with high fidelity, including characters, objects, and scenes.

Exquisite demand exists for customizing the pretrained large text-to-image model, $e.g.$ Stable Diffusion, to generate innovative concepts, such as the users themselves. However, the newly-added concept from previous customization methods often shows weaker combination abilities than the original ones even given several images during training. We thus propose a new personalization method that allows for the seamless integration of a unique individual into the pre-trained diffusion model using just $one\ facial\ photograph$ and only $1024\ learnable\ parameters$ under $3\ minutes$. So we can effortlessly generate stunning images of this person in any pose or position, interacting with anyone and doing anything imaginable from text prompts. To achieve this, we first analyze and build a well-defined celeb basis from the embedding space of the pre-trained large text encoder. Then, given one facial photo as the target identity, we generate its own embedding by optimizing the weight of this basis and locking all other parameters. Empowered by the proposed celeb basis, the new identity in our customized model showcases a better concept combination ability than previous personalization methods. Besides, our model can also learn several new identities at once and interact with each other where the previous customization model fails to. Project page is at: http://celeb-basis.github.io. Code is at: https://github.com/ygtxr1997/CelebBasis.

Accurate Story visualization requires several necessary elements, such as identity consistency across frames, the alignment between plain text and visual content, and a reasonable layout of objects in images. Most previous works endeavor to meet these requirements by fitting a text-to-image (T2I) model on a set of videos in the same style and with the same characters, e.g., the FlintstonesSV dataset. However, the learned T2I models typically struggle to adapt to new characters, scenes, and styles, and often lack the flexibility to revise the layout of the synthesized images. This paper proposes a system for generic interactive story visualization, capable of handling multiple novel characters and supporting the editing of layout and local structure. It is developed by leveraging the prior knowledge of large language and T2I models, trained on massive corpora. The system comprises four interconnected components: story-to-prompt generation (S2P), text-to-layout generation (T2L), controllable text-to-image generation (C-T2I), and image-to-video animation (I2V). First, the S2P module converts concise story information into detailed prompts required for subsequent stages. Next, T2L generates diverse and reasonable layouts based on the prompts, offering users the ability to adjust and refine the layout to their preferences. The core component, C-T2I, enables the creation of images guided by layouts, sketches, and actor-specific identifiers to maintain consistency and detail across visualizations. Finally, I2V enriches the visualization process by animating the generated images. Extensive experiments and a user study are conducted to validate the effectiveness and flexibility of interactive editing of the proposed system.

We show that pre-trained Generative Adversarial Networks (GANs) such as StyleGAN and BigGAN can be used as a latent bank to improve the performance of image super-resolution. While most existing perceptual-oriented approaches attempt to generate realistic outputs through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass for restoration. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Employing priors from different generative models allows GLEAN to be applied to diverse categories (e.g., human faces, cats, buildings, and cars). We further present a lightweight version of GLEAN, named LightGLEAN, which retains only the critical components in GLEAN. Notably, LightGLEAN consists of only 21% of parameters and 35% of FLOPs while achieving comparable image quality. We extend our method to different tasks including image colorization and blind image restoration, and extensive experiments show that our proposed models perform favorably in comparison to existing methods. Codes and models are available at https://github.com/open-mmlab/mmediting.

Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image or video by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g.scale and rotation) and the resolution gap (e.g.HR and LR). To tackle these challenges, we propose$C^{2}$-Matching in this work, which performs explicit robust matching crossing transformation and resolution. 1) To bridge the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) To address the resolution gap, we adopt teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue between input images and reference images. In addition, to faithfully evaluate the performance of Reference-based Image Super-Resolution (Ref Image SR) under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. We also extend$C^{2}$-Matching to Reference-based Video Super-Resolution (Ref VSR) task, where an image taken in a similar scene serves as the HR reference image. Extensive experiments demonstrate that our proposed$C^{2}$-Matching significantly outperforms state of the arts by up to 0.7dB on the standard CUFED5 benchmark and also boosts the performance of video super-resolution by incorporating the$C^{2}$-Matching component into Video SR pipelines. Notably,$C^{2}$-Matching also shows great generalizability on WR-SR dataset as well as robustness across large scale and rotation transformations. Codes and datasets are available athttps://github.com/yumingj/C2-Matching.

This paper explores training efficient VGG-style super-resolution (SR) networks with the structural re-parameterization technique. The general pipeline of re-parameterization is to train networks with multi-branch topology first, and then merge them into standard 3x3 convolutions for efficient inference. In this work, we revisit those primary designs and investigate essential components for re-parameterizing SR networks. First of all, we find that batch normalization (BN) is important to bring training non-linearity and improve the final performance. However, BN is typically ignored in SR, as it usually degrades the performance and introduces unpleasant artifacts. We carefully analyze the cause of BN issue and then propose a straightforward yet effective solution. In particular, we first train SR networks with mini-batch statistics as usual, and then switch to using population statistics at the later training period. While we have successfully re-introduced BN into SR, we further design a new re-parameterizable block tailored for SR, namely RepSR. It consists of a clean residual path and two expand-and-squeeze convolution paths with the modified BN. Extensive experiments demonstrate that our simple RepSR is capable of achieving superior performance to previous SR re-parameterization methods among different model sizes. In addition, our RepSR can achieve a better trade-off between performance and actual running time (throughput) than previous SR methods. Codes are available at https://github.com/TencentARC/RepSR.

Compositing portrait photographs or videos to novel backgrounds is an important application in computational photography. Seamless blending along boundaries and globally harmonic colors are two desired properties of the photo-realistic composition of foregrounds and new backgrounds. Existing works are dedicated to either foreground alpha matte generation or after-blending harmonization, leading to sub-optimal background replacement when putting foregrounds and backgrounds together. In this work, we unify the two objectives in a single framework to obtain realistic portrait image composites. Specifically, we investigate the usage of a target background and find that a complete background plays a vital role in both seamlessly blending and harmonization. We develop a network to learn the composition process given an imperfect alpha matte with appearance features extracted from the complete background to adjust color distribution. Our dedicated usage of a complete background enables realistic portrait image composition and also temporally stable results on videos. Extensive quantitative and qualitative experiments on both synthetic and real-world data demonstrate that our method achieves state-of-the-art performance.

The alignment of adjacent frames is considered an essential operation in video super-resolution (VSR). Advanced VSR models, including the latest VSR Transformers, are generally equipped with well-designed alignment modules. However, the progress of the self-attention mechanism may violate this common sense. In this paper, we rethink the role of alignment in VSR Transformers and make several counter-intuitive observations. Our experiments show that: (i) VSR Transformers can directly utilize multi-frame information from unaligned videos, and (ii) existing alignment methods are sometimes harmful to VSR Transformers. These observations indicate that we can further improve the performance of VSR Transformers simply by removing the alignment module and adopting a larger attention window. Nevertheless, such designs will dramatically increase the computational burden, and cannot deal with large motions. Therefore, we propose a new and efficient alignment method called patch alignment, which aligns image patches instead of pixels. VSR Transformers equipped with patch alignment could demonstrate state-of-the-art performance on multiple benchmarks. Our work provides valuable insights on how multi-frame information is used in VSR and how to select alignment methods for different networks/datasets. Codes and models will be released at https://github.com/XPixelGroup/RethinkVSRAlignment.

This paper studies the problem of real-world video super-resolution (VSR) for animation videos, and reveals three key improvements for practical animation VSR. First, recent real-world super-resolution approaches typically rely on degradation simulation using basic operators without any learning capability, such as blur, noise, and compression. In this work, we propose to learn such basic operators from real low-quality animation videos, and incorporate the learned ones into the degradation generation pipeline. Such neural-network-based basic operators could help to better capture the distribution of real degradations. Second, a large-scale high-quality animation video dataset, AVC, is built to facilitate comprehensive training and evaluations for animation VSR. Third, we further investigate an efficient multi-scale network structure. It takes advantage of the efficiency of unidirectional recurrent networks and the effectiveness of sliding-window-based methods. Thanks to the above delicate designs, our method, AnimeSR, is capable of restoring real-world low-quality animation videos effectively and efficiently, achieving superior performance to previous state-of-the-art methods.

Very deep Convolutional Neural Networks (CNNs) have greatly improved the performance on various image restoration tasks. However, this comes at a price of increasing computational burden, hence limiting their practical usages. We observe that some corrupted image regions are inherently easier to restore than others since the distortion and content vary within an image. To leverage this, we propose Path-Restore, a multi-path CNN with a pathfinder that can dynamically select an appropriate route for each image region. We train the pathfinder using reinforcement learning with a difficulty-regulated reward. This reward is related to the performance, complexity and "the difficulty of restoring a region". A policy mask is further investigated to jointly process all the image regions. We conduct experiments on denoising and mixed restoration tasks. The results show that our method achieves comparable or superior performance to existing approaches with less computational cost. In particular, Path-Restore is effective for real-world denoising, where the noise distribution varies across different regions on a single image. Compared to the state-of-the-art RIDNet [1], our method achieves comparable performance and runs 2.7x faster on the realistic Darmstadt Noise Dataset [2]. Models and codes are available on the project page: https://www.mmlab-ntu.com/project/pathrestore/.

Deformable convolution, originally proposed for the adaptation to geometric variations of objects, has recently shown compelling performance in aligning multiple frames and is increasingly adopted for video super-resolution. Despite its remarkable performance, its underlying mechanism for alignment remains unclear. In this study, we carefully investigate the relation between deformable alignment and the classic flow-based alignment. We show that deformable convolution can be decomposed into a combination of spatial warping and convolution. This decomposition reveals the commonality of deformable alignment and flow-based alignment in formulation, but with a key difference in their offset diversity. We further demonstrate through experiments that the increased diversity in deformable alignment yields better-aligned features, and hence significantly improves the quality of video super-resolution output. Based on our observations, we propose an offset-fidelity loss that guides the offset learning with optical flow. Experiments show that our loss successfully avoids the overflow of offsets and alleviates the instability problem of deformable alignment. Aside from the contributions to deformable alignment, our formulation inspires a more flexible approach to introduce offset diversity to flow-based alignment, improving its performance.

Reference-based Super-Resolution (Ref-SR) has recently emerged as a promising paradigm to enhance a low-resolution (LR) input image by introducing an additional high-resolution (HR) reference image. Existing Ref-SR methods mostly rely on implicit correspondence matching to borrow HR textures from reference images to compensate for the information loss in input images. However, performing local transfer is difficult because of two gaps between input and reference images: the transformation gap (e.g. scale and rotation) and the resolution gap (e.g. HR and LR). To tackle these challenges, we propose C2-Matching in this work, which produces explicit robust matching crossing transformation and resolution. 1) For the transformation gap, we propose a contrastive correspondence network, which learns transformation-robust correspondences using augmented views of the input image. 2) For the resolution gap, we adopt a teacher-student correlation distillation, which distills knowledge from the easier HR-HR matching to guide the more ambiguous LR-HR matching. 3) Finally, we design a dynamic aggregation module to address the potential misalignment issue. In addition, to faithfully evaluate the performance of Ref-SR under a realistic setting, we contribute the Webly-Referenced SR (WR-SR) dataset, mimicking the practical usage scenario. Extensive experiments demonstrate that our proposed C2-Matching significantly outperforms state of the arts by over 1dB on the standard CUFED5 benchmark. Notably, it also shows great generalizability on WR-SR dataset as well as robustness across large scale and rotation transformations1.

We show that pre-trained Generative Adversarial Networks (GANs), e.g., StyleGAN, can be used as a latent bank to improve the restoration quality of large-factor image super-resolution (SR). While most existing SR approaches attempt to generate realistic textures through learning with adversarial loss, our method, Generative LatEnt bANk (GLEAN), goes beyond existing practices by directly leveraging rich and diverse priors encapsulated in a pre-trained GAN. But unlike prevalent GAN inversion methods that require expensive image-specific optimization at runtime, our approach only needs a single forward pass to generate the upscaled image. GLEAN can be easily incorporated in a simple encoder-bank-decoder architecture with multi-resolution skip connections. Switching the bank allows the method to deal with images from diverse categories, e.g., cat, building, human face, and car. Images upscaled by GLEAN show clear improvements in terms of fidelity and texture faithfulness in comparison to existing methods as shown in Fig. 1.

Video super-resolution (VSR) approaches tend to have more components than the image counterparts as they need to exploit the additional temporal dimension. Complex designs are not uncommon. In this study, we wish to untangle the knots and reconsider some most essential components for VSR guided by four basic functionalities, i.e., Propagation, Alignment, Aggregation, and Upsampling. By reusing some existing components added with minimal redesigns, we show a succinct pipeline, BasicVSR, that achieves appealing improvements in terms of speed and restoration quality in comparison to many state-of-the-art algorithms. We conduct systematic analysis to explain how such gain can be obtained and discuss the pitfalls. We further show the extensibility of BasicVSR by presenting an information-refill mechanism and a coupled propagation scheme to facilitate information aggregation. The BasicVSR and its extension, IconVSR, can serve as strong baselines for future VSR approaches.

Blind face restoration usually relies on facial priors, such as facial geometry prior or reference prior, to restore realistic and faithful details. However, very low-quality inputs cannot offer accurate geometric prior while high-quality references are inaccessible, limiting the applicability in real-world scenarios. In this work, we propose GFP-GAN that leverages rich and diverse priors encapsulated in a pretrained face GAN for blind face restoration. This Generative Facial Prior (GFP) is incorporated into the face restoration process via spatial feature transform layers, which allow our method to achieve a good balance of realness and fidelity. Thanks to the powerful generative facial prior and delicate designs, our GFP-GAN could jointly restore facial details and enhance colors with just a single forward pass, while GAN inversion methods require image-specific optimization at inference. Extensive experiments show that our method achieves superior performance to prior art on both synthetic and real-world datasets.

SinGAN shows impressive capability in learning internal patch distribution despite its limited effective receptive field. We are interested in knowing how such a translationinvariant convolutional generator could capture the global structure with just a spatially i.i.d. input. In this work, taking SinGAN and StyleGAN2 as examples, we show that such capability, to a large extent, is brought by the implicit positional encoding when using zero padding in the generators. Such positional encoding is indispensable for generating images with high fidelity. The same phenomenon is observed in other generative architectures such as DCGAN and PGGAN. We further show that zero padding leads to an unbalanced spatial bias with a vague relation between locations. To offer a better spatial inductive bias, we investigate alternative positional encodings and analyze their effects. Based on a more flexible positional encoding explicitly, we propose a new multi-scale training strategy and demonstrate its effectiveness in the state-of-the-art unconditional generator StyleGAN2. Besides, the explicit spatial inductive bias substantially improves SinGAN for more versatile image manipulation.1

Colorization has attracted increasing interest in recent years. Classic reference-based methods usually rely on external color images for plausible results. A large image database or online search engine is inevitably required for retrieving such exemplars. Recent deep-learning-based methods could automatically colorize images at a low cost. However, unsatisfactory artifacts and incoherent colors are always accompanied. In this work, we aim at recovering vivid colors by leveraging the rich and diverse color priors encapsulated in a pretrained Generative Adversarial Networks (GAN). Specifically, we first "retrieve" matched features (similar to exemplars) via a GAN encoder and then incorporate these features into the colorization process with feature modulations. Thanks to the powerful generative color prior and delicate designs, our method could produce vivid colors with a single forward pass. Moreover, it is highly convenient to obtain diverse results by modifying GAN latent codes. Our method also inherits the merit of interpretable controls of GANs and could attain controllable and smooth transitions by walking through GAN latent space. Extensive experiments and user studies demonstrate that our method achieves superior performance than previous works.

Recent blind super-resolution (SR) methods typically consist of two branches, one for degradation prediction and the other for conditional restoration. However, our experiments show that a one-branch network can achieve comparable performance to the two-branch scheme. Then we wonder: how can one-branch networks automatically learn to distinguish degradations? To find the answer, we propose a new diagnostic tool -- Filter Attribution method based on Integral Gradient (FAIG). Unlike previous integral gradient methods, our FAIG aims at finding the most discriminative filters instead of input pixels/features for degradation removal in blind SR networks. With the discovered filters, we further develop a simple yet effective method to predict the degradation of an input image. Based on FAIG, we show that, in one-branch blind SR networks, 1) we could find a very small number of (1%) discriminative filters for each specific degradation; 2) The weights, locations and connections of the discovered filters are all important to determine the specific network function. 3) The task of degradation prediction can be implicitly realized by these discriminative filters without explicit supervised learning. Our findings can not only help us better understand network behaviors inside one-branch blind SR networks, but also provide guidance on designing more efficient architectures and diagnosing networks for blind SR.

Deep convolutional neural network has demonstrated its capability of learning a deterministic mapping for the desired imagery effect. However, the large variety of user flavors motivates the possibility of continuous transition among different output effects. Unlike existing methods that require a specific design to achieve one particular transition (e.g., style transfer), we propose a simple yet universal approach to attain a smooth control of diverse imagery effects in many low-level vision tasks, including image restoration, image-to-image translation, and style transfer. Specifically, our method, namely Deep Network Interpolation (DNI), applies linear interpolation in the parameter space of two or more correlated networks. A smooth control of imagery effects can be achieved by tweaking the interpolation coefficients. In addition to DNI and its broad applications, we also investigate the mechanism of network interpolation from the perspective of learned filters.

Despite that convolutional neural networks (CNN) have recently demonstrated high-quality reconstruction for single-image super-resolution (SR), recovering natural and realistic texture remains a challenging problem. In this paper, we show that it is possible to recover textures faithful to semantic classes. In particular, we only need to modulate features of a few intermediate layers in a single network conditioned on semantic segmentation probability maps. This is made possible through a novel Spatial Feature Transform (SFT) layer that generates affine transformation parameters for spatial-wise feature modulation. SFT layers can be trained end-to-end together with the SR network using the same loss function. During testing, it accepts an input image of arbitrary size and generates a high-resolution image with just a single forward pass conditioned on the categorical priors. Our final results show that an SR network equipped with SFT can generate more realistic and visually pleasing textures in comparison to state-of-the-art SRGAN [27] and EnhanceNet [38].