Daily papers related to Image/Video/Multimodal Generation from cs.CV
April 04, 2026
Conditional image editing aims to modify a source image according to textual prompts and optional reference guidance. Such editing is crucial in scenarios requiring strict structural control (i.e., anomaly insertion in driving scenes and complex human pose transformation). Despite recent advances in large-scale editing models (i.e., Seedream, Nano Banana, etc), most approaches rely on single-step generation. This paradigm often lacks explicit quality control, may introduce excessive deviation from the original image, and frequently produces structural artifacts or environment-inconsistent modifications, typically requiring manual prompt tuning to achieve acceptable results. We propose \textbf{CAMEO}, a structured multi-agent framework that reformulates conditional editing as a quality-aware, feedback-driven process rather than a one-shot generation task. CAMEO decomposes editing into coordinated stages of planning, structured prompting, hypothesis generation, and adaptive reference grounding, where external guidance is invoked only when task complexity requires it. To overcome the lack of intrinsic quality control in existing methods, evaluation is embedded directly within the editing loop. Intermediate results are iteratively refined through structured feedback, forming a closed-loop process that progressively corrects structural and contextual inconsistencies. We evaluate CAMEO on anomaly insertion and human pose switching tasks. Across multiple strong editing backbones and independent evaluation models, CAMEO consistently achieves 20\% more win rate on average compared to multiple state-of-the-art models, demonstrating improved robustness, controllability, and structural reliability in conditional image editing.
TLDR: The paper introduces CAMEO, a multi-agent framework for conditional image editing that uses a feedback-driven process to improve quality, controllability, and structural reliability compared to single-step generation methods.
TLDR: 该论文介绍了一种名为CAMEO的多智能体条件图像编辑框架，该框架采用反馈驱动流程，与单步生成方法相比，提高了质量、可控性和结构可靠性。
Read Paper (PDF)Distilling video generation models to extremely low inference budgets (e.g., 2--4 NFEs) is crucial for real-time deployment, yet remains challenging. Trajectory-style consistency distillation often becomes conservative under complex video dynamics, yielding an over-smoothed appearance and weak motion. Distribution matching distillation (DMD) can recover sharp, mode-seeking samples, but its local training signals do not explicitly regularize how denoising updates compose across timesteps, making composed rollouts prone to drift. To overcome this challenge, we propose Self-Consistent Distribution Matching Distillation (SC-DMD), which explicitly regularizes the endpoint-consistent composition of consecutive denoising updates. For real-time autoregressive video generation, we further treat the KV cache as a quality parameterized condition and propose Cache-Distribution-Aware training. This training scheme applies SC-DMD over multi-step rollouts and introduces a cache-conditioned feature alignment objective that steers low-quality outputs toward high-quality references. Across extensive experiments on both non-autoregressive backbones (e.g., Wan~2.1) and autoregressive real-time paradigms (e.g., Self Forcing), our method, dubbed \textbf{Salt}, consistently improves low-NFE video generation quality while remaining compatible with diverse KV-cache memory mechanisms. Source code will be released at \href{https://github.com/XingtongGe/Salt}{https://github.com/XingtongGe/Salt}.
TLDR: The paper introduces Self-Consistent Distribution Matching Distillation (SC-DMD) and cache-aware training for fast, high-quality video generation, especially at low inference budgets.
TLDR: 该论文介绍了一种名为自洽分布匹配蒸馏 (SC-DMD) 的方法和缓存感知训练，用于快速生成高质量视频，尤其是在低推理预算下。
Read Paper (PDF)Autoregressive (AR) video diffusion models enable long-form video generation but remain expensive due to repeated multi-step denoising. Existing training-free acceleration methods rely on binary cache-or-recompute decisions, overlooking intermediate cases where direct reuse is too coarse yet full recomputation is unnecessary. Moreover, asynchronous AR schedules assign different noise levels to co-generated frames, yet existing methods process the entire valid interval uniformly. To address these AR-specific inefficiencies, we present SCOPE, a training-free framework for efficient AR video diffusion. SCOPE introduces a tri-modal scheduler over cache, predict, and recompute, where prediction via noise-level Taylor extrapolation fills the gap between reuse and recomputation with explicit stability controls backed by error propagation analysis. It further introduces selective computation that restricts execution to the active frame interval. On MAGI-1 and SkyReels-V2, SCOPE achieves up to 4.73x speedup while maintaining quality comparable to the original output, outperforming all training-free baselines.
TLDR: The paper introduces SCOPE, a training-free framework for accelerating autoregressive video diffusion models using a tri-modal scheduler (cache, predict, recompute) and selective computation, achieving significant speedups while maintaining video quality.
TLDR: 该论文介绍了SCOPE，一个无需训练的框架，通过使用三模态调度器（缓存、预测、重新计算）和选择性计算来加速自回归视频扩散模型，从而在保持视频质量的同时实现显著的加速。
Read Paper (PDF)Creating flexible 3D scenes from a single image is vital when direct 3D data acquisition is costly or impractical. We introduce NavCrafter, a novel framework that explores 3D scenes from a single image by synthesizing novel-view video sequences with camera controllability and temporal-spatial consistency. NavCrafter leverages video diffusion models to capture rich 3D priors and adopts a geometry-aware expansion strategy to progressively extend scene coverage. To enable controllable multi-view synthesis, we introduce a multi-stage camera control mechanism that conditions diffusion models with diverse trajectories via dual-branch camera injection and attention modulation. We further propose a collision-aware camera trajectory planner and an enhanced 3D Gaussian Splatting (3DGS) pipeline with depth-aligned supervision, structural regularization and refinement. Extensive experiments demonstrate that NavCrafter achieves state-of-the-art novel-view synthesis under large viewpoint shifts and substantially improves 3D reconstruction fidelity.
TLDR: NavCrafter is a framework for exploring and reconstructing 3D scenes from a single image by synthesizing novel-view video sequences using video diffusion models and geometry-aware scene expansion.
TLDR: NavCrafter是一个框架，通过使用视频扩散模型和几何感知的场景扩展，从单张图像探索和重建3D场景，并合成新视角的视频序列。
Read Paper (PDF)Despite advancements in generating visually stunning content, video diffusion models (VDMs) often yield physically inconsistent results due to pixel-only reconstruction. To address this, we propose MMPhysVideo, the first framework to scale physical plausibility in video generation through joint multimodal modeling. We recast perceptual cues, specifically semantics, geometry, and spatio-temporal trajectory, into a unified pseudo-RGB format, enabling VDMs to directly capture complex physical dynamics. To mitigate cross-modal interference, we propose a Bidirectionally Controlled Teacher architecture, which utilizes parallel branches to fully decouple RGB and perception processing and adopts two zero-initialized control links to gradually learn pixel-wise consistency. For inference efficiency, the teacher's physical prior is distilled into a single-stream student model via representation alignment. Furthermore, we present MMPhysPipe, a scalable data curation and annotation pipeline tailored for constructing physics-rich multimodal datasets. MMPhysPipe employs a vision-language model (VLM) guided by a chain-of-visual-evidence rule to pinpoint physical subjects, enabling expert models to extract multi-granular perceptual information. Without additional inference costs, MMPhysVideo consistently improves physical plausibility and visual quality over advanced models across various benchmarks and achieves state-of-the-art performance compared to existing methods.
TLDR: The paper introduces MMPhysVideo, a framework that improves physical plausibility in video generation by incorporating perceptual cues as pseudo-RGB channels and using a Bidirectionally Controlled Teacher architecture, along with a data curation pipeline called MMPhysPipe.
TLDR: 本文介绍了MMPhysVideo，该框架通过将感知线索整合为伪RGB通道，并使用双向控制的教师架构，以及一个名为MMPhysPipe的数据标注流程，从而提高视频生成中物理真实性。
Read Paper (PDF)Text-driven human motion generation aims to synthesize realistic motion sequences that follow textual descriptions. Despite recent advances, accurately aligning motion dynamics with textual semantics remains a fundamental challenge. In this paper, we revisit text-to-motion generation from the perspective of motion-language alignment and propose MLA-Gen, a framework that integrates global motion priors with fine-grained local conditioning. This design enables the model to capture common motion patterns, while establishing detailed alignment between texts and motions. Furthermore, we identify a previously overlooked attention sink phenomenon in human motion generation, where attention disproportionately concentrates on the start text token, limiting the utilization of informative textual cues and leading to degraded semantic grounding. To analyze this issue, we introduce SinkRatio, a metric for measuring attention concentration, and develop alignment-aware masking and control strategies to regulate attention during generation. Extensive experiments demonstrate that our approach consistently improves both motion quality and motion-language alignment over strong baselines. Code will be released upon acceptance.
TLDR: This paper introduces MLA-Gen, a framework for text-driven human motion generation that focuses on improving motion-language alignment by addressing the attention sink phenomenon and integrating global motion priors with fine-grained local conditioning. They also introduce a metric called SinkRatio.
TLDR: 本文介绍了MLA-Gen，一个用于文本驱动的人体运动生成框架，通过解决注意力沉没现象，并将全局运动先验与细粒度局部条件结合，来提高运动-语言对齐性。他们还引入了一个名为SinkRatio的指标。
Read Paper (PDF)Editing animatable human avatars typically relies on sparse supervision, often a few edited keyframes, yet naively fitting a reconstructed avatar to these edits frequently causes identity leakage and pose-dependent temporal flicker. We argue that these failures are best understood as an ill-conditioned inversion: the available edited constraints do not sufficiently determine the latent directions responsible for the intended edit. We propose a conditioning-guided edited reconstruction framework that performs editing as a constrained inversion in a structured avatar latent space, restricting updates to a low-dimensional, part-specific edit subspace to prevent unintended identity changes. Crucially, we design the editing constraints during inversion by optimizing a conditioning objective derived from a local linearization of the full decoding-and-rendering pipeline, yielding an edit-subspace information matrix whose spectrum predicts stability and drives frame reweighting / keyframe activation. The resulting method operates on small subspace matrices and can be implemented efficiently (e.g., via Hessian-vector products), and improves stability under limited edited supervision.
TLDR: This paper introduces a method for stable avatar editing from sparse supervision using information-regularized constrained inversion in a structured latent space, focusing on preventing identity leakage and temporal flicker.
TLDR: 本文提出了一种从稀疏监督中进行稳定头像编辑的方法，该方法在结构化的潜在空间中使用信息正则化约束反演，重点是防止身份泄露和时间闪烁。
Read Paper (PDF)The generation of 3D hand-object interactions (HOIs) from text is crucial for dexterous robotic grasping and VR/AR content generation, requiring both high visual fidelity and physical plausibility. Nevertheless, the ill-posed problem of mesh extraction from text-generated Gaussians, and physics-based optimization on the erroneous meshes pose challenges. To address these issues, we introduce THOM, a training-free framework that generates photorealistic, physically plausible 3D HOI meshes without the need for a template object mesh. THOM employs a two-stage pipeline, initially generating the hand and object Gaussians, followed by physics-based HOI optimization. Our new mesh extraction method and vertex-to-Gaussian mapping explicitly assign Gaussian elements to mesh vertices, allowing topology-aware regularization. Furthermore, we improve the physical plausibility of interactions by VLM-guided translation refinement and contact-aware optimization. Comprehensive experiments demonstrate that THOM consistently surpasses state-of-the-art methods in terms of text alignment, visual realism, and interaction plausibility.
TLDR: THOM is a training-free framework that generates photorealistic and physically plausible 3D hand-object interaction meshes from text by using a two-stage pipeline involving Gaussian generation and physics-based optimization, outperforming existing methods in text alignment, visual realism, and interaction plausibility.
TLDR: THOM是一个无需训练的框架，通过一个两阶段流程，即高斯生成和基于物理的优化，从文本生成逼真且物理上合理的3D手-物体交互网格，并在文本对齐、视觉逼真度和交互合理性方面优于现有方法。
Read Paper (PDF)Deploying large multimodal language models at scale is constrained by token-based inference costs, yet the cost-performance behavior of visual prompting strategies remains poorly characterized. We introduce Image Prompt Packaging (IPPg), a prompting paradigm that embeds structured text directly into images to reduce text token overhead, and benchmark it across five datasets, three frontier models (GPT-4.1, GPT-4o, Claude 3.5 Sonnet), and two task families (VQA and code generation). We derive a cost formulation decomposing savings by token type and show IPPg achieves 35.8--91.0\% inference cost reductions. Despite token compression of up to 96\%, accuracy remains competitive in many settings, though outcomes are highly model- and task-dependent: GPT-4.1 achieves simultaneous accuracy and cost gains on CoSQL, while Claude 3.5 incurs cost increases on several VQA benchmarks. Systematic error analysis yields a failure-mode taxonomy: spatial reasoning, non-English inputs, and character-sensitive operations are most vulnerable, while schema-structured tasks benefit most. A 125-configuration rendering ablation reveals accuracy shifts of 10--30 percentage points, establishing visual encoding choices as a first-class variable in multimodal system design.
TLDR: The paper introduces Image Prompt Packaging (IPPg), a method to reduce token costs in large multimodal models by embedding structured text into images, achieving significant cost reductions while maintaining competitive accuracy in certain tasks, though performance varies across models and tasks.
TLDR: 该论文介绍了图像提示打包（IPPg），一种通过将结构化文本嵌入图像来降低大型多模态模型中token成本的方法。IPPg在某些任务中实现了显著的成本降低，同时保持了具有竞争力的精度，但性能因模型和任务而异。
Read Paper (PDF)The scarcity of labeled satellite imagery remains a fundamental bottleneck for deep-learning (DL)-based wildfire monitoring systems. This paper investigates whether a diffusion-based foundation model for Earth Observation (EO), EarthSynth, can synthesize realistic post-wildfire Sentinel-2 RGB imagery conditioned on existing burn masks, without task-specific retraining. Using burn masks derived from the CalFireSeg-50 dataset (Martin et al., 2025), we design and evaluate six controlled experimental configurations that systematically vary: (i) pipeline architecture (mask-only full generation vs. inpainting with pre-fire context), (ii) prompt engineering strategy (three hand-crafted prompts and a VLM-generated prompt via Qwen2-VL), and (iii) a region-wise color-matching post-processing step. Quantitative assessment on 10 stratified test samples uses four complementary metrics: Burn IoU, burn-region color distance (ΔC_burn), Darkness Contrast, and Spectral Plausibility. Results show that inpainting-based pipelines consistently outperform full-tile generation across all metrics, with the structured inpainting prompt achieving the best spatial alignment (Burn IoU = 0.456) and burn saliency (Darkness Contrast = 20.44), while color matching produces the lowest color distance (ΔC_burn = 63.22) at the cost of reduced burn saliency. VLM-assisted inpainting is competitive with hand-crafted prompts. These findings provide a foundation for incorporating generative data augmentation into wildfire detection pipelines. Code and experiments are available at: https://www.kaggle.com/code/valeriamartinh/genai-all-runned
TLDR: This paper explores using a diffusion model (EarthSynth) to generate synthetic satellite imagery of post-wildfire scenes conditioned on burn masks, finding that inpainting-based pipelines with color matching produce the most realistic results and VLM prompts can be competitive with hand-engineered prompts.
TLDR: 本文探讨了使用扩散模型（EarthSynth）生成以燃烧掩码为条件的火灾后场景合成卫星图像，发现基于修复的流水线与颜色匹配产生了最真实的结果，并且VLM prompt可以与手工设计的prompt相媲美。
Read Paper (PDF)Multi-agent trajectory generation in team sports requires models that capture both the diversity of possible plays and realistic spatial coordination between players on plays. Standard generative approaches such as Conditional Variational Autoencoders (CVAE) and diffusion models struggle with this task, exhibiting posterior collapse or convergence to the dataset mean. Moreover, most trajectory prediction methods operate in a forecasting regime that requires multiple frames of observed history, limiting their use for play design where only the initial formation is available. We present PlayGen-MoG, an extensible framework for formation-conditioned play generation that addresses these challenges through three design choices: 1/ a Mixture-of-Gaussians (MoG) output head with shared mixture weights across all agents, where a single set of weights selects a play scenario that couples all players' trajectories, 2/ relative spatial attention that encodes pairwise player positions and distances as learned attention biases, and 3/ non-autoregressive prediction of absolute displacements from the initial formation, eliminating cumulative error drift and removing the dependence on observed trajectory history, enabling realistic play generation from a single static formation alone. On American football tracking data, PlayGen-MoG achieves 1.68 yard ADE and 3.98 yard FDE while maintaining full utilization of all 8 mixture components with entropy of 2.06 out of 2.08, and qualitatively confirming diverse generation without mode collapse.
TLDR: The paper introduces PlayGen-MoG, a framework for diverse multi-agent trajectory generation in team sports using a Mixture-of-Gaussians approach and relative spatial attention, enabling play generation from static formations without relying on observed trajectory history.
TLDR: 该论文介绍了PlayGen-MoG，一个用于团队运动中多样化多智能体轨迹生成的框架，它使用混合高斯方法和相对空间注意力，能够仅从静态阵型生成战术，而无需依赖观察到的轨迹历史。
Read Paper (PDF)