Inspired by DeepSeek-R1's success in eliciting reasoning abilities through rule-based reinforcement learning (RL), we introduce Video-R1 as the first attempt to systematically explore the R1 paradigm for incentivizing video reasoning within multimodal large language models (MLLMs). However, directly applying RL training with the GRPO algorithm to video reasoning presents two primary challenges: (i) a lack of temporal modeling for video reasoning, and (ii) the scarcity of high-quality video-reasoning data. To address these issues, we first propose the T-GRPO algorithm, which encourages models to utilize temporal information in videos for reasoning. Additionally, instead of relying solely on video data, we incorporate high-quality image-reasoning data into the training process. We have constructed two datasets: Video-R1-CoT-165k for SFT cold start and Video-R1-260k for RL training, both comprising image and video data. Experimental results demonstrate that Video-R1 achieves significant improvements on video reasoning benchmarks such as VideoMMMU and VSI-Bench, as well as on general video benchmarks including MVBench and TempCompass, etc. Notably, Video-R1-7B attains a 37.1% accuracy
We introduce Reangle-A-Video, a unified framework for generating synchronized multi-view videos from a single input video. Unlike mainstream approaches that train multi-view video diffusion models on large-scale 4D datasets, our method reframes the multi-view video generation task as video-to-videos translation, leveraging publicly available image and video diffusion priors. In essence, Reangle-A-Video operates in two stages. (1) Multi-View Motion Learning: An image-to-video diffusion transformer is synchronously fine-tuned in a self-supervised manner to distill view-invariant motion from a set of warped videos. (2) Multi-View Consistent Image-to-Images Translation: The first frame of the input video is warped and inpainted into various camera perspectives under an inference-time cross-view consistency guidance using DUSt3R, generating multi-view consistent starting images. Extensive experiments on static view transport and dynamic camera control show that Reangle-A-Video surpasses existing methods, establishing a new solution for multi-view video generation. We will publicly release our code and data. Project page: https://hyeonho99.github.io/reangle-a-video/
A unified video and action model holds significant promise for robotics, where videos provide rich scene information for action prediction, and actions provide dynamics information for video prediction. However, effectively combining video generation and action prediction remains challenging, and current video generation-based methods struggle to match the performance of direct policy learning in action accuracy and inference speed. To bridge this gap, we introduce the Unified Video Action model (UVA), which jointly optimizes video and action predictions to achieve both high accuracy and efficient action inference. The key lies in learning a joint video-action latent representation and decoupling video-action decoding. The joint latent representation bridges the visual and action domains, effectively modeling the relationship between video and action sequences. Meanwhile, the decoupled decoding, powered by two lightweight diffusion heads, enables high-speed action inference by bypassing video generation during inference. Such a unified framework further enables versatile functionality through masked input training. By selectively masking actions or videos, a single model can tackle
Video generation models have demonstrated great capabilities of producing impressive monocular videos, however, the generation of 3D stereoscopic video remains under-explored. We propose a pose-free and training-free approach for generating 3D stereoscopic videos using an off-the-shelf monocular video generation model. Our method warps a generated monocular video into camera views on stereoscopic baseline using estimated video depth, and employs a novel frame matrix video inpainting framework. The framework leverages the video generation model to inpaint frames observed from different timestamps and views. This effective approach generates consistent and semantically coherent stereoscopic videos without scene optimization or model fine-tuning. Moreover, we develop a disocclusion boundary re-injection scheme that further improves the quality of video inpainting by alleviating the negative effects propagated from disoccluded areas in the latent space. We validate the efficacy of our proposed method by conducting experiments on videos from various generative models, including Sora [4 ], Lumiere [2], WALT [8 ], and Zeroscope [ 42]. The experiments demonstrate that our method has a sign
Recently, breakthroughs in video modeling have allowed for controllable camera trajectories in generated videos. However, these methods cannot be directly applied to user-provided videos that are not generated by a video model. In this paper, we present ReCapture, a method for generating new videos with novel camera trajectories from a single user-provided video. Our method allows us to re-generate the reference video, with all its existing scene motion, from vastly different angles and with cinematic camera motion. Notably, using our method we can also plausibly hallucinate parts of the scene that were not observable in the reference video. Our method works by (1) generating a noisy anchor video with a new camera trajectory using multiview diffusion models or depth-based point cloud rendering and then (2) regenerating the anchor video into a clean and temporally consistent reangled video using our proposed masked video fine-tuning technique.
Recent advancements in video large language models (Video LLMs) have significantly advanced the field of video question answering (VideoQA). While existing methods perform well on short videos, they often struggle with long-range reasoning in longer videos. To scale Video LLMs for longer video content, frame sub-sampling (selecting frames at regular intervals) is commonly used. However, this approach is suboptimal, often leading to the loss of crucial frames or the inclusion of redundant information from multiple similar frames. Missing key frames impairs the model's ability to answer questions accurately, while redundant frames lead the model to focus on irrelevant video segments and increase computational resource consumption. In this paper, we investigate the use of a general-purpose text-to-video moment retrieval model to guide the frame sampling process. We propose "moment sampling", a novel, model-agnostic approach that enables the model to select the most relevant frames according to the context of the question. Specifically, we employ a lightweight moment retrieval model to prioritize frame selection. By focusing on the frames most pertinent to the given question, our metho
Adaptive streaming of segmented video over HTTP typically relies on a predefined set of bitrate-resolution pairs, known as a bitrate ladder. However, fixed ladders often overlook variations in content and decoding complexities, leading to suboptimal trade-offs between encoding time, decoding efficiency, and video quality. This article introduces DQ-Ladder, a deep reinforcement learning (DRL)-based scheme for constructing time- and quality-aware bitrate ladders for adaptive video streaming applications. DQ-Ladder employs predicted decoding time, quality scores, and bitrate levels per segment as inputs to a Deep Q-Network (DQN) agent, guided by a weighted reward function of decoding time, video quality, and resolution smoothness. We leverage machine learning models to predict decoding time, bitrate level, and objective quality metrics (VMAF, XPSNR), eliminating the need for exhaustive encoding or quality metric computation. We evaluate DQ-Ladder using the Versatile Video Coding (VVC) toolchain (VVenC/VVdeC) on 750 video sequences across six Apple HLS-compliant resolutions and 41 quantization parameters. Experimental results against four baselines show that DQ-Ladder achieves BD-rate
Video tokenizers are essential for latent video diffusion models, converting raw video data into spatiotemporally compressed latent spaces for efficient training. However, extending state-of-the-art video tokenizers to achieve a temporal compression ratio beyond 4x without increasing channel capacity poses significant challenges. In this work, we propose an alternative approach to enhance temporal compression. We find that the reconstruction quality of temporally subsampled videos from a low-compression encoder surpasses that of high-compression encoders applied to original videos. This indicates that high-compression models can leverage representations from lower-compression models. Building on this insight, we develop a bootstrapped high-temporal-compression model that progressively trains high-compression blocks atop well-trained lower-compression models. Our method includes a cross-level feature-mixing module to retain information from the pretrained low-compression model and guide higher-compression blocks to capture the remaining details from the full video sequence. Evaluation of video benchmarks shows that our method significantly improves reconstruction quality while incre
Human processes video reasoning in a sequential spatio-temporal reasoning logic, we first identify the relevant frames ("when") and then analyse the spatial relationships ("where") between key objects, and finally leverage these relationships to draw inferences ("what"). However, can Video Large Language Models (Video-LLMs) also "reason through a sequential spatio-temporal logic" in videos? Existing Video-LLM benchmarks primarily focus on assessing object presence, neglecting relational reasoning. Consequently, it is difficult to measure whether a model truly comprehends object interactions (actions/events) in videos or merely relies on pre-trained "memory" of co-occurrences as biases in generating answers. In this work, we introduce a Video Spatio-Temporal Reasoning (V-STaR) benchmark to address these shortcomings. The key idea is to decompose video understanding into a Reverse Spatio-Temporal Reasoning (RSTR) task that simultaneously evaluates what objects are present, when events occur, and where they are located while capturing the underlying Chain-of-thought (CoT) logic. To support this evaluation, we construct a dataset to elicit the spatial-temporal reasoning process of Vide
While language models have become impactful in many real-world applications, video generation remains largely confined to entertainment. Motivated by video's inherent capacity to demonstrate physical-world information that is difficult to convey through language alone (e.g., imagine teaching someone to tie a tie using only text), we identify an underutilized opportunity to extend video as a new answer modality for Next-Event Prediction (NEP), formalized as Video-Next-Event Prediction (VNEP). While the established NEP task takes a video with a procedural or predictive question as input to predict the next event in text, VNEP requires dynamic video responses. This shift from telling to showing unlocks more intuitive and customized answers for procedural learning and creative exploration. However, this task remains challenging for existing models, as it demands an understanding of multimodal input, instruction-conditioned reasoning, and the generation of video with visual and semantic consistency. To address this, we introduce VANS, a model that leverages reinforcement learning to align a Vision-Language Model (VLM) with a Video Diffusion Model (VDM) for VNEP. The core of VANS is our
Unified, generalizable semantic control in video generation remains a critical open challenge. Existing methods either introduce artifacts by enforcing inappropriate pixel-wise priors from structure-based controls, or rely on non-generalizable, condition-specific finetuning or task-specific architectures. We introduce Video-As-Prompt (VAP), a new paradigm that reframes this problem as in-context generation. VAP leverages a reference video as a direct semantic prompt, guiding a frozen Video Diffusion Transformer (DiT) via a plug-and-play Mixture-of-Transformers (MoT) expert. This architecture prevents catastrophic forgetting and is guided by a temporally biased position embedding that eliminates spurious mapping priors for robust context retrieval. To power this approach and catalyze future research, we built VAP-Data, the largest dataset for semantic-controlled video generation with over 100K paired videos across 100 semantic conditions. As a single unified model, VAP sets a new state-of-the-art for open-source methods, achieving a 38.7% user preference rate that rivals leading condition-specific commercial models. VAP's strong zero-shot generalization and support for various downs
Text-to-video diffusion models have made remarkable advancements. Driven by their ability to generate temporally coherent videos, research on zero-shot video editing using these fundamental models has expanded rapidly. To enhance editing quality, structural controls are frequently employed in video editing. Among these techniques, cross-attention mask control stands out for its effectiveness and efficiency. However, when cross-attention masks are naively applied to video editing, they can introduce artifacts such as blurring and flickering. Our experiments uncover a critical factor overlooked in previous video editing research: cross-attention masks are not consistently clear but vary with model structure and denoising timestep. To address this issue, we propose the metric Mask Matching Cost (MMC) that quantifies this variability and propose FreeMask, a method for selecting optimal masks tailored to specific video editing tasks. Using MMC-selected masks, we further improve the masked fusion mechanism within comprehensive attention features, e.g., temp, cross, and self-attention modules. Our approach can be seamlessly integrated into existing zero-shot video editing frameworks with
We present Imagen Video, a text-conditional video generation system based on a cascade of video diffusion models. Given a text prompt, Imagen Video generates high definition videos using a base video generation model and a sequence of interleaved spatial and temporal video super-resolution models. We describe how we scale up the system as a high definition text-to-video model including design decisions such as the choice of fully-convolutional temporal and spatial super-resolution models at certain resolutions, and the choice of the v-parameterization of diffusion models. In addition, we confirm and transfer findings from previous work on diffusion-based image generation to the video generation setting. Finally, we apply progressive distillation to our video models with classifier-free guidance for fast, high quality sampling. We find Imagen Video not only capable of generating videos of high fidelity, but also having a high degree of controllability and world knowledge, including the ability to generate diverse videos and text animations in various artistic styles and with 3D object understanding. See https://imagen.research.google/video/ for samples.
Video compression is a standard procedure applied to all videos to minimize storage and transmission demands while preserving visual quality as much as possible. Therefore, evaluating the visual quality of compressed videos is crucial for guiding the practical usage and further development of video compression algorithms. Although numerous compressed video quality assessment (VQA) methods have been proposed, they often lack the generalization capability needed to handle the increasing diversity of video types, particularly high dynamic range (HDR) content. In this paper, we introduce CompressedVQA-HDR, an effective VQA framework designed to address the challenges of HDR video quality assessment. Specifically, we adopt the Swin Transformer and SigLip 2 as the backbone networks for the proposed full-reference (FR) and no-reference (NR) VQA models, respectively. For the FR model, we compute deep structural and textural similarities between reference and distorted frames using intermediate-layer features extracted from the Swin Transformer as its quality-aware feature representation. For the NR model, we extract the global mean of the final-layer feature maps from SigLip 2 as its quali
With the scale capability of increasing training data, model size, and computational cost, video generation has achieved impressive results in digital creation, enabling users to express creativity across various domains. Recently, researchers in Large Language Models (LLMs) have expanded the scaling to test-time, which can significantly improve LLM performance by using more inference-time computation. Instead of scaling up video foundation models through expensive training costs, we explore the power of Test-Time Scaling (TTS) in video generation, aiming to answer the question: if a video generation model is allowed to use non-trivial amount of inference-time compute, how much can it improve generation quality given a challenging text prompt. In this work, we reinterpret the test-time scaling of video generation as a searching problem to sample better trajectories from Gaussian noise space to the target video distribution. Specifically, we build the search space with test-time verifiers to provide feedback and heuristic algorithms to guide searching process. Given a text prompt, we first explore an intuitive linear search strategy by increasing noise candidates at inference time.
Video coding has traditionally been developed to support services such as video streaming, videoconferencing, digital TV, and so on. The main intent was to enable human viewing of the encoded content. However, with the advances in deep neural networks (DNNs), encoded video is increasingly being used for automatic video analytics performed by machines. In applications such as automatic traffic monitoring, analytics such as vehicle detection, tracking and counting, would run continuously, while human viewing could be required occasionally to review potential incidents. To support such applications, a new paradigm for video coding is needed that will facilitate efficient representation and compression of video for both machine and human use in a scalable manner. In this manuscript, we introduce an end-to-end learnable video codec that supports a machine vision task in its base layer, while its enhancement layer, together with the base layer, supports input reconstruction for human viewing. The proposed system is constructed based on the concept of conditional coding to achieve better compression gains. Comprehensive experimental evaluations conducted on four standard video datasets de
We are interested in enabling visual planning for complex long-horizon tasks in the space of generated videos and language, leveraging recent advances in large generative models pretrained on Internet-scale data. To this end, we present video language planning (VLP), an algorithm that consists of a tree search procedure, where we train (i) vision-language models to serve as both policies and value functions, and (ii) text-to-video models as dynamics models. VLP takes as input a long-horizon task instruction and current image observation, and outputs a long video plan that provides detailed multimodal (video and language) specifications that describe how to complete the final task. VLP scales with increasing computation budget where more computation time results in improved video plans, and is able to synthesize long-horizon video plans across different robotics domains: from multi-object rearrangement, to multi-camera bi-arm dexterous manipulation. Generated video plans can be translated into real robot actions via goal-conditioned policies, conditioned on each intermediate frame of the generated video. Experiments show that VLP substantially improves long-horizon task success rate
Preparing high-quality 360-degree video for HTTP Adaptive Streaming requires encoding each sequence into multiple representations spanning different resolutions and quantization parameters (QPs). For ultra-high-resolution immersive content such as 8K 360-degree video, this process is computationally intensive due to the large number of representations and the high complexity of modern codecs. This paper investigates fast multirate encoding strategies that reduce encoding time by reusing encoder analysis information across QPs and resolutions. We evaluate two cross-resolution information-reuse pipelines that differ in how reference encodes propagate across resolutions: (i) a strict HD -> 4K -> 8K cascade with scaled analysis reuse, and (ii) a resolution-anchored scheme that initializes each resolution with its own highest-bitrate reference before guiding dependent encodes. In addition to evaluating these pipelines on standard equirectangular projection content, we also apply the same two pipelines to cubemap-projection (CMP) tiling, where each 360-degree frame is partitioned into independently encoded tiles. CMP introduces substantial parallelism, while still benefiting from t
During the past decade, the Unmanned-Aerial-Vehicles (UAVs) have attracted increasing attention due to their flexible, extensive, and dynamic space-sensing capabilities. The volume of video captured by UAVs is exponentially growing along with the increased bitrate generated by the advancement of the sensors mounted on UAVs, bringing new challenges for on-device UAV storage and air-ground data transmission. Most existing video compression schemes were designed for natural scenes without consideration of specific texture and view characteristics of UAV videos. In this work, we first contribute a detailed analysis of the current state of the field of UAV video coding. Then we propose to establish a novel task for learned UAV video coding and construct a comprehensive and systematic benchmark for such a task, present a thorough review of high quality UAV video datasets and benchmarks, and contribute extensive rate-distortion efficiency comparison of learned and conventional codecs after. Finally, we discuss the challenges of encoding UAV videos. It is expected that the benchmark will accelerate the research and development in video coding on drone platforms.
Video Question Answering (VideoQA) represents a crucial intersection between video understanding and language processing, requiring both discriminative unimodal comprehension and sophisticated cross-modal interaction for accurate inference. Despite advancements in multi-modal pre-trained models and video-language foundation models, these systems often struggle with domain-specific VideoQA due to their generalized pre-training objectives. Addressing this gap necessitates bridging the divide between broad cross-modal knowledge and the specific inference demands of VideoQA tasks. To this end, we introduce HeurVidQA, a framework that leverages domain-specific entity-action heuristics to refine pre-trained video-language foundation models. Our approach treats these models as implicit knowledge engines, employing domain-specific entity-action prompters to direct the model's focus toward precise cues that enhance reasoning. By delivering fine-grained heuristics, we improve the model's ability to identify and interpret key entities and actions, thereby enhancing its reasoning capabilities. Extensive evaluations across multiple VideoQA datasets demonstrate that our method significantly outp