PyTorchKR Papers

최신 논문

총 1,879편
태그·기간 등 상세 필터

카테고리 탐색

로그인 후 이 태그 구독

arXiv:2405.16728

Towards Multi-Task Multi-Modal Models: A Video Generative Perspective

Advancements in language foundation models have primarily fueled the recent surge in artificial intelligence. In contrast, generative learning of non-textual modalities, especially videos, significantly trails behind language modeling. This thesis chronicles our endeavor to build multi-task models for generating videos and other modalities under diverse conditions, as well as for understanding and compression applications. Given the high dimensionality of visual data, we pursue concise and accurate latent representations. Our video-native spatial-temporal tokenizers preserve high fidelity. We unveil a novel approach to mapping bidirectionally between visual observation and interpretable lexical terms. Furthermore, our scalable visual token representation proves beneficial across generation, compression, and understanding tasks. This achievement marks the first instances of language models surpassing diffusion models in visual synthesis and a video tokenizer outperforming industry-standard codecs. Within these multi-modal latent spaces, we study the design of multi-task generative models. Our masked multi-task transformer excels at the quality, efficiency, and flexibility of video generation. We enable a frozen language model, trained solely on text, to generate visual content. Finally, we build a scalable generative multi-modal transformer trained from scratch, enabling the generation of videos containing high-fidelity motion with the corresponding audio given diverse conditions. Throughout the course, we have shown the effectiveness of integrating multiple tasks, crafting high-fidelity latent representation, and generating multiple modalities. This work suggests intriguing potential for future exploration in generating non-textual data and enabling real-time, interactive experiences across various media forms.

arXiv:2405.17398공식 코드 · Papers with Code

Vista: A Generalizable Driving World Model with High Fidelity and Versatile Controllability

World models can foresee the outcomes of different actions, which is of paramount importance for autonomous driving. Nevertheless, existing driving world models still have limitations in generalization to unseen environments, prediction fidelity of critical details, and action controllability for flexible application. In this paper, we present Vista, a generalizable driving world model with high fidelity and versatile controllability. Based on a systematic diagnosis of existing methods, we introduce several key ingredients to address these limitations. To accurately predict real-world dynamics at high resolution, we propose two novel losses to promote the learning of moving instances and structural information. We also devise an effective latent replacement approach to inject historical frames as priors for coherent long-horizon rollouts. For action controllability, we incorporate a versatile set of controls from high-level intentions (command, goal point) to low-level maneuvers (trajectory, angle, and speed) through an efficient learning strategy. After large-scale training, the capabilities of Vista can seamlessly generalize to different scenarios. Extensive experiments on multiple datasets show that Vista outperforms the most advanced general-purpose video generator in over 70% of comparisons and surpasses the best-performing driving world model by 55% in FID and 27% in FVD. Moreover, for the first time, we utilize the capacity of Vista itself to establish a generalizable reward for real-world action evaluation without accessing the ground truth actions.

arXiv:2405.15406공식 코드 · Papers with Code

A Misleading Gallery of Fluid Motion by Generative Artificial Intelligence

In this technical report, we extensively investigate the accuracy of outputs from well-known generative artificial intelligence (AI) applications in response to prompts describing common fluid motion phenomena familiar to the fluid mechanics community. We examine a range of applications, including Midjourney, Dall-E, Runway ML, Microsoft Designer, Gemini, Meta AI, and Leonardo AI, introduced by prominent companies such as Google, OpenAI, Meta, and Microsoft. Our text prompts for generating images or videos include examples such as "Von Karman vortex street", "flow past an airfoil", "Kelvin-Helmholtz instability", "shock waves on a sharp-nosed supersonic body", etc. We compare the images generated by these applications with real images from laboratory experiments and numerical software. Our findings indicate that these generative AI models are not adequately trained in fluid dynamics imagery, leading to potentially misleading outputs. Beyond text-to-image/video generation, we further explore the transition from image/video to text generation using these AI tools, aiming to investigate the accuracy of their descriptions of fluid motion phenomena. This report serves as a cautionary note for educators in academic institutions, highlighting the potential for these tools to mislead students. It also aims to inform researchers at these renowned companies, encouraging them to address this issue. We conjecture that a primary reason for this shortcoming is the limited access to copyright-protected fluid motion images from scientific journals.

arXiv:2405.13865

ReVideo: Remake a Video with Motion and Content Control

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.

arXiv:2405.14864

Video Diffusion Models are Training-free Motion Interpreter and Controller

Video generation primarily aims to model authentic and customized motion across frames, making understanding and controlling the motion a crucial topic. Most diffusion-based studies on video motion focus on motion customization with training-based paradigms, which, however, demands substantial training resources and necessitates retraining for diverse models. Crucially, these approaches do not explore how video diffusion models encode cross-frame motion information in their features, lacking interpretability and transparency in their effectiveness. To answer this question, this paper introduces a novel perspective to understand, localize, and manipulate motion-aware features in video diffusion models. Through analysis using Principal Component Analysis (PCA), our work discloses that robust motion-aware feature already exists in video diffusion models. We present a new MOtion FeaTure (MOFT) by eliminating content correlation information and filtering motion channels. MOFT provides a distinct set of benefits, including the ability to encode comprehensive motion information with clear interpretability, extraction without the need for training, and generalizability across diverse architectures. Leveraging MOFT, we propose a novel training-free video motion control framework. Our method demonstrates competitive performance in generating natural and faithful motion, providing architecture-agnostic insights and applicability in a variety of downstream tasks.

arXiv:2405.13557공식 코드 · Papers with Code

MotionCraft: Physics-based Zero-Shot Video Generation

Generating videos with realistic and physically plausible motion is one of the main recent challenges in computer vision. While diffusion models are achieving compelling results in image generation, video diffusion models are limited by heavy training and huge models, resulting in videos that are still biased to the training dataset. In this work we propose MotionCraft, a new zero-shot video generator to craft physics-based and realistic videos. MotionCraft is able to warp the noise latent space of an image diffusion model, such as Stable Diffusion, by applying an optical flow derived from a physics simulation. We show that warping the noise latent space results in coherent application of the desired motion while allowing the model to generate missing elements consistent with the scene evolution, which would otherwise result in artefacts or missing content if the flow was applied in the pixel space. We compare our method with the state-of-the-art Text2Video-Zero reporting qualitative and quantitative improvements, demonstrating the effectiveness of our approach to generate videos with finely-prescribed complex motion dynamics. Project page: https://mezzelfo.github.io/MotionCraft/

arXiv:2405.13195

CamViG: Camera Aware Image-to-Video Generation with Multimodal Transformers

We extend multimodal transformers to include 3D camera motion as a conditioning signal for the task of video generation. Generative video models are becoming increasingly powerful, thus focusing research efforts on methods of controlling the output of such models. We propose to add virtual 3D camera controls to generative video methods by conditioning generated video on an encoding of three-dimensional camera movement over the course of the generated video. Results demonstrate that we are (1) able to successfully control the camera during video generation, starting from a single frame and a camera signal, and (2) we demonstrate the accuracy of the generated 3D camera paths using traditional computer vision methods.

arXiv:2405.14475

MagicDrive3D: Controllable 3D Generation for Any-View Rendering in Street Scenes

While controllable generative models for images and videos have achieved remarkable success, high-quality models for 3D scenes, particularly in unbounded scenarios like autonomous driving, remain underdeveloped due to high data acquisition costs. In this paper, we introduce MagicDrive3D, a novel pipeline for controllable 3D street scene generation that supports multi-condition control, including BEV maps, 3D objects, and text descriptions. Unlike previous methods that reconstruct before training the generative models, MagicDrive3D first trains a video generation model and then reconstructs from the generated data. This innovative approach enables easily controllable generation and static scene acquisition, resulting in high-quality scene reconstruction. To address the minor errors in generated content, we propose deformable Gaussian splatting with monocular depth initialization and appearance modeling to manage exposure discrepancies across viewpoints. Validated on the nuScenes dataset, MagicDrive3D generates diverse, high-quality 3D driving scenes that support any-view rendering and enhance downstream tasks like BEV segmentation. Our results demonstrate the framework's superior performance, showcasing its potential for autonomous driving simulation and beyond.

arXiv:2405.14664

Fisher Flow Matching for Generative Modeling over Discrete Data

Generative modeling over discrete data has recently seen numerous success stories, with applications spanning language modeling, biological sequence design, and graph-structured molecular data. The predominant generative modeling paradigm for discrete data is still autoregressive, with more recent alternatives based on diffusion or flow-matching falling short of their impressive performance in continuous data settings, such as image or video generation. In this work, we introduce Fisher-Flow, a novel flow-matching model for discrete data. Fisher-Flow takes a manifestly geometric perspective by considering categorical distributions over discrete data as points residing on a statistical manifold equipped with its natural Riemannian metric: the $\textit{Fisher-Rao metric}$. As a result, we demonstrate discrete data itself can be continuously reparameterised to points on the positive orthant of the $d$-hypersphere $\mathbb{S}^d_+$, which allows us to define flows that map any source distribution to target in a principled manner by transporting mass along (closed-form) geodesics of $\mathbb{S}^d_+$. Furthermore, the learned flows in Fisher-Flow can be further bootstrapped by leveraging Riemannian optimal transport leading to improved training dynamics. We prove that the gradient flow induced by Fisher-Flow is optimal in reducing the forward KL divergence. We evaluate Fisher-Flow on an array of synthetic and diverse real-world benchmarks, including designing DNA Promoter, and DNA Enhancer sequences. Empirically, we find that Fisher-Flow improves over prior diffusion and flow-matching models on these benchmarks.

arXiv:2405.14582

PoseCrafter: One-Shot Personalized Video Synthesis Following Flexible Pose Control

In this paper, we introduce PoseCrafter, a one-shot method for personalized video generation following the control of flexible poses. Built upon Stable Diffusion and ControlNet, we carefully design an inference process to produce high-quality videos without the corresponding ground-truth frames. First, we select an appropriate reference frame from the training video and invert it to initialize all latent variables for generation. Then, we insert the corresponding training pose into the target pose sequences to enhance faithfulness through a trained temporal attention module. Furthermore, to alleviate the face and hand degradation resulting from discrepancies between poses of training videos and inference poses, we implement simple latent editing through an affine transformation matrix involving facial and hand landmarks. Extensive experiments on several datasets demonstrate that PoseCrafter achieves superior results to baselines pre-trained on a vast collection of videos under 8 commonly used metrics. Besides, PoseCrafter can follow poses from different individuals or artificial edits and simultaneously retain the human identity in an open-domain training video. Our project page is available at https://ml-gsai.github.io/PoseCrafter-demo/.

arXiv:2405.12843공식 코드 · Papers with Code

OpenCarbonEval: A Unified Carbon Emission Estimation Framework in Large-Scale AI Models

In recent years, large-scale auto-regressive models have made significant progress in various tasks, such as text or video generation. However, the environmental impact of these models has been largely overlooked, with a lack of assessment and analysis of their carbon footprint. To address this gap, we introduce OpenCarbonEval, a unified framework for integrating large-scale models across diverse modalities to predict carbon emissions, which could provide AI service providers and users with a means to estimate emissions beforehand and help mitigate the environmental pressure associated with these models. In OpenCarbonEval, we propose a dynamic throughput modeling approach that could capture workload and hardware fluctuations in the training process for more precise emissions estimates. Our evaluation results demonstrate that OpenCarbonEval can more accurately predict training emissions than previous methods, and can be seamlessly applied to different modal tasks. Specifically, we show that OpenCarbonEval achieves superior performance in predicting carbon emissions for both visual models and language models. By promoting sustainable AI development and deployment, OpenCarbonEval can help reduce the environmental impact of large-scale models and contribute to a more environmentally responsible future for the AI community.

arXiv:2405.12796

DisenStudio: Customized Multi-subject Text-to-Video Generation with Disentangled Spatial Control

Generating customized content in videos has received increasing attention recently. However, existing works primarily focus on customized text-to-video generation for single subject, suffering from subject-missing and attribute-binding problems when the video is expected to contain multiple subjects. Furthermore, existing models struggle to assign the desired actions to the corresponding subjects (action-binding problem), failing to achieve satisfactory multi-subject generation performance. To tackle the problems, in this paper, we propose DisenStudio, a novel framework that can generate text-guided videos for customized multiple subjects, given few images for each subject. Specifically, DisenStudio enhances a pretrained diffusion-based text-to-video model with our proposed spatial-disentangled cross-attention mechanism to associate each subject with the desired action. Then the model is customized for the multiple subjects with the proposed motion-preserved disentangled finetuning, which involves three tuning strategies: multi-subject co-occurrence tuning, masked single-subject tuning, and multi-subject motion-preserved tuning. The first two strategies guarantee the subject occurrence and preserve their visual attributes, and the third strategy helps the model maintain the temporal motion-generation ability when finetuning on static images. We conduct extensive experiments to demonstrate our proposed DisenStudio significantly outperforms existing methods in various metrics. Additionally, we show that DisenStudio can be used as a powerful tool for various controllable generation applications.

arXiv:2405.11473공식 코드 · Papers with Code

FIFO-Diffusion: Generating Infinite Videos from Text without Training

We propose a novel inference technique based on a pretrained diffusion model for text-conditional video generation. Our approach, called FIFO-Diffusion, is conceptually capable of generating infinitely long videos without additional training. This is achieved by iteratively performing diagonal denoising, which simultaneously processes a series of consecutive frames with increasing noise levels in a queue; our method dequeues a fully denoised frame at the head while enqueuing a new random noise frame at the tail. However, diagonal denoising is a double-edged sword as the frames near the tail can take advantage of cleaner frames by forward reference but such a strategy induces the discrepancy between training and inference. Hence, we introduce latent partitioning to reduce the training-inference gap and lookahead denoising to leverage the benefit of forward referencing. Practically, FIFO-Diffusion consumes a constant amount of memory regardless of the target video length given a baseline model, while well-suited for parallel inference on multiple GPUs. We have demonstrated the promising results and effectiveness of the proposed methods on existing text-to-video generation baselines. Generated video examples and source codes are available at our project page.

arXiv:2405.10674공식 코드 · Papers with Code

From Sora What We Can See: A Survey of Text-to-Video Generation

With impressive achievements made, artificial intelligence is on the path forward to artificial general intelligence. Sora, developed by OpenAI, which is capable of minute-level world-simulative abilities can be considered as a milestone on this developmental path. However, despite its notable successes, Sora still encounters various obstacles that need to be resolved. In this survey, we embark from the perspective of disassembling Sora in text-to-video generation, and conducting a comprehensive review of literature, trying to answer the question, \textit{From Sora What We Can See}. Specifically, after basic preliminaries regarding the general algorithms are introduced, the literature is categorized from three mutually perpendicular dimensions: evolutionary generators, excellent pursuit, and realistic panorama. Subsequently, the widely used datasets and metrics are organized in detail. Last but more importantly, we identify several challenges and open problems in this domain and propose potential future directions for research and development.

arXiv:2405.09266

Dance Any Beat: Blending Beats with Visuals in Dance Video Generation

Generating dance from music is crucial for advancing automated choreography. Current methods typically produce skeleton keypoint sequences instead of dance videos and lack the capability to make specific individuals dance, which reduces their real-world applicability. These methods also require precise keypoint annotations, complicating data collection and limiting the use of self-collected video datasets. To overcome these challenges, we introduce a novel task: generating dance videos directly from images of individuals guided by music. This task enables the dance generation of specific individuals without requiring keypoint annotations, making it more versatile and applicable to various situations. Our solution, the Dance Any Beat Diffusion model (DabFusion), utilizes a reference image and a music piece to generate dance videos featuring various dance types and choreographies. The music is analyzed by our specially designed music encoder, which identifies essential features including dance style, movement, and rhythm. DabFusion excels in generating dance videos not only for individuals in the training dataset but also for any previously unseen person. This versatility stems from its approach of generating latent optical flow, which contains all necessary motion information to animate any person in the image. We evaluate DabFusion's performance using the AIST++ dataset, focusing on video quality, audio-video synchronization, and motion-music alignment. We propose a 2D Motion-Music Alignment Score (2D-MM Align), which builds on the Beat Alignment Score to more effectively evaluate motion-music alignment for this new task. Experiments show that our DabFusion establishes a solid baseline for this innovative task. Video results can be found on our project page: https://DabFusion.github.io.

arXiv:2405.08720

The Lost Melody: Empirical Observations on Text-to-Video Generation From A Storytelling Perspective

Text-to-video generation task has witnessed a notable progress, with the generated outcomes reflecting the text prompts with high fidelity and impressive visual qualities. However, current text-to-video generation models are invariably focused on conveying the visual elements of a single scene, and have so far been indifferent to another important potential of the medium, namely a storytelling. In this paper, we examine text-to-video generation from a storytelling perspective, which has been hardly investigated, and make empirical remarks that spotlight the limitations of current text-to-video generation scheme. We also propose an evaluation framework for storytelling aspects of videos, and discuss the potential future directions.

arXiv:2405.06547공식 코드 · Papers with Code

OneTo3D: One Image to Re-editable Dynamic 3D Model and Video Generation

One image to editable dynamic 3D model and video generation is novel direction and change in the research area of single image to 3D representation or 3D reconstruction of image. Gaussian Splatting has demonstrated its advantages in implicit 3D reconstruction, compared with the original Neural Radiance Fields. As the rapid development of technologies and principles, people tried to used the Stable Diffusion models to generate targeted models with text instructions. However, using the normal implicit machine learning methods is hard to gain the precise motions and actions control, further more, it is difficult to generate a long content and semantic continuous 3D video. To address this issue, we propose the OneTo3D, a method and theory to used one single image to generate the editable 3D model and generate the targeted semantic continuous time-unlimited 3D video. We used a normal basic Gaussian Splatting model to generate the 3D model from a single image, which requires less volume of video memory and computer calculation ability. Subsequently, we designed an automatic generation and self-adaptive binding mechanism for the object armature. Combined with the re-editable motions and actions analyzing and controlling algorithm we proposed, we can achieve a better performance than the SOTA projects in the area of building the 3D model precise motions and actions control, and generating a stable semantic continuous time-unlimited 3D video with the input text instructions. Here we will analyze the detailed implementation methods and theories analyses. Relative comparisons and conclusions will be presented. The project code is open source.

arXiv:2405.05039

Reviewing Intelligent Cinematography: AI research for camera-based video production

This paper offers the first comprehensive review of artificial intelligence (AI) research in the context of real camera content acquisition for entertainment purposes and is aimed at both researchers and cinematographers. Addressing the lack of review papers in the field of intelligent cinematography} (IC) and the breadth of related computer vision research, we present a holistic view of the IC landscape while providing technical insight, important for experts across disciplines. We provide technical background on generative AI, object detection, automated camera calibration and 3-D content acquisition, with references to assist non-technical readers. The application sections categorize work in terms of four production types: General Production, Virtual Production, Live Production and Aerial Production. Within each application section, we (1) sub-classify work according to research topic and (2) describe the trends and challenges relevant to each type of production. In the final chapter, we address the greater scope of IC research and summarize the significant potential of this area to influence the creative industries sector. We suggest that work relating to virtual production has the greatest potential to impact other mediums of production, driven by the growing interest in LED volumes/stages for in-camera virtual effects (ICVFX) and automated 3-D capture for virtual modeling of real world scenes and actors. We also address ethical and legal concerns regarding the use of creative AI that impact on artists, actors, technologists and the general public.

arXiv:2405.04682공식 코드 · Papers with Code

TALC: Time-Aligned Captions for Multi-Scene Text-to-Video Generation

Most of these text-to-video (T2V) generative models often produce single-scene video clips that depict an entity performing a particular action (e.g., 'a red panda climbing a tree'). However, it is pertinent to generate multi-scene videos since they are ubiquitous in the real-world (e.g., 'a red panda climbing a tree' followed by 'the red panda sleeps on the top of the tree'). To generate multi-scene videos from the pretrained T2V model, we introduce a simple and effective Time-Aligned Captions (TALC) framework. Specifically, we enhance the text-conditioning mechanism in the T2V architecture to recognize the temporal alignment between the video scenes and scene descriptions. For instance, we condition the visual features of the earlier and later scenes of the generated video with the representations of the first scene description (e.g., 'a red panda climbing a tree') and second scene description (e.g., 'the red panda sleeps on the top of the tree'), respectively. As a result, we show that the T2V model can generate multi-scene videos that adhere to the multi-scene text descriptions and be visually consistent (e.g., entity and background). Further, we finetune the pretrained T2V model with multi-scene video-text data using the TALC framework. We show that the TALC-finetuned model outperforms the baseline by achieving a relative gain of 29% in the overall score, which averages visual consistency and text adherence using human evaluation.

arXiv:2405.04133

Exposing AI-generated Videos: A Benchmark Dataset and a Local-and-Global Temporal Defect Based Detection Method

The generative model has made significant advancements in the creation of realistic videos, which causes security issues. However, this emerging risk has not been adequately addressed due to the absence of a benchmark dataset for AI-generated videos. In this paper, we first construct a video dataset using advanced diffusion-based video generation algorithms with various semantic contents. Besides, typical video lossy operations over network transmission are adopted to generate degraded samples. Then, by analyzing local and global temporal defects of current AI-generated videos, a novel detection framework by adaptively learning local motion information and global appearance variation is constructed to expose fake videos. Finally, experiments are conducted to evaluate the generalization and robustness of different spatial and temporal domain detection methods, where the results can serve as the baseline and demonstrate the research challenge for future studies.