CVPR 2026 Papers

Recent progress in 4D representations, such as Dynamic NeRF and 4D Gaussian Splatting (4DGS), has enabled dynamic 4D scene reconstruction. However, text-driven 4D scene editing remains under-explored due to the challenge of ensuring both multi-view and temporal consistency across space and time during editing.Existing studies rely on 2D diffusion models that edit frames independently, often causing motion distortion, geometric drift, and incomplete editing. We introduce Dynamic-eDiTor, a training-free text-driven 4D editing framework leveraging Multimodal Diffusion Transformer (MM-DiT) and 4DGS. This mechanism consists of Spatio-Temporal Sub-Grid Attention (STGA) for locally consistent cross-view and temporal fusion, and Context Token Propagation (CTP) for global propagation via token inheritance and optical-flow-guided token replacement. Together, these components allow Dynamic-eDiTor to perform seamless, globally consistent multi-view video without additional training and directly optimize pre-trained source 4DGS.Extensive experiments on multi-view video dataset DyNeRF demonstrate that our method achieves superior editing fidelity and both multi-view and temporal consistency prior approaches. Project page for results and code: https://di-lee.github.io/dynamic-eDiTor

Diffusion models degrade images through noise, and reversing this process reveals an information hierarchy across timesteps. Scale-space theory exhibits a similar hierarchy via low-pass filtering. We formalize this connection and show that highly noisy diffusion states contain no more information than small, downsampled images - raising the question of why they must be processed at full resolution. To address this, we fuse scale spaces into the diffusion process by formulating a family of diffusion models with generalized linear degradations and practical implementations. Using downsampling as the degradation yields our proposed \ours. To support \ours, we introduce \ourmodel , a UNet variant that performs resolution-preserving and resolution-increasing denoising using only the necessary parts of the network. We evaluate our framework on CelebA and ImageNet and analyze its scaling behavior across resolutions and network depths. Our website is available at prateksha.github.io/projects/scale-space-diffusion.

Woven fabric materials are widely used in rendering applications, yet designing realistic examples typically involves multiple stages, requiring expertise in weaving principles and texture authoring. Recent advances have explored diffusion models to streamline this process; however, pre-trained diffusion models often struggle to generate intricate yarn-level details that conform to weaving rules. To address this, we present FabricGen, an end-to-end framework for generating high-quality woven fabric materials from textual descriptions. A key insight of our method is the decomposition of macro-scale textures and micro-scale weaving patterns. To generate macro-scale textures free from microstructures, we fine-tune pre-trained diffusion models on a collected dataset of microstructure-free fabrics. As for micro-scale weaving patterns, we develop an enhanced procedural geometric model capable of synthesizing natural yarn-level geometry with yarn sliding and flyaway fibers. The procedural model is driven by a specialized large language model, WeavingLLM, which is fine-tuned on an annotated dataset of formatted weaving drafts, and prompt-tuned with domain-specific fabric expertise. Through fine-tuning and prompt tuning, WeavingLLM learns to design weaving drafts and fabric parameters from textual prompts, enabling the procedural model to produce diverse weaving patterns that stick to weaving principles. The generated macro-scale texture, along with the micro-scale geometry, can be used for fabric rendering. Consequently, our framework produces materials with significantly richer detail and realism compared to prior generative models.

Generative pose priors have recently emerged as a powerful tool for inference under occlusion or noise. Yet today's strongest generative paradigm, flow matching, remains unused for human pose due to two fundamental barriers: the absence of a pre-trained flow prior and the non-Euclidean nature of articulated poses. We overcome both by introducing PoseD-Flow, a novel framework to unify Riemannian Flow Matching (RFM) with training-free guidance for 3D human pose recovery. PoseD-Flow is composed of two contributions: (i) PoseRFM, the first RFM model of human pose, defined directly on the product manifold of joint rotations, and (ii) Riemannian D-Flow, a principled guidance mechanism that, by differentiating through its ODE sampling dynamics, conditions PoseRFM at inference without any task-specific training. Our theoretical analysis shows that the induced dynamics are shaped by data covariance and manifold curvature, yielding a bias toward realistic poses. Across pose completion, denoising, and inverse kinematics, PoseD-Flow establishes new state of the art, particularly under noise, occlusion, and partial observations.

Diffusion model (DM) based Video Super-Resolution (VSR) approaches achieve impressive perceptual quality. Diffusion model (DM) based Video Super-Resolution (VSR) approaches achieve impressive perceptual quality. However, existing DM-based VSR methods over-prioritize perceptual synthesis while neglecting fidelity gains from accurate alignment and sufficient compensation. In this paper, within the DM-based VSR pipeline, we revisit the role of alignment and compensation between adjacent video frames and reveal two crucial observations: (a) the feature domain is better suited than the pixel domain for information compensation due to its stronger spatial and temporal correlations, and (b) warping at an upscaled resolution better preserves high-frequency information, but this benefit is not necessarily monotonic. Therefore, we propose a novel Densely Guided diffusion model with Aligned Features for Video Super-Resolution (DGAF-VSR), with an Optical Guided Warping Module (OGWM) to maintain high-frequency details in the aligned features and a Feature-wise Temporal Condition Module (FTCM) to deliver dense guidance in the feature domain. Extensive experiments on synthetic and real-world datasets demonstrate that DGAF-VSR surpasses state-of-the-art methods in key aspects of VSR, including perceptual quality (35.82% DISTS reduction), fidelity (0.20 dB PSNR gain), and temporal consistency (30.37% tLPIPS reduction).

Test-time alignment (TTA) aims to adapt models to specific rewards during inference. However, existing methods tend to either under-optimise or over-optimise (reward hack) the target reward function. We propose Null-Text Test-Time Alignment (Null-TTA), which aligns diffusion models by optimising the unconditional embedding in classifier-free guidance, rather than manipulating latent or noise variables. Due to the structured semantic nature of the text embedding space, this ensures alignment occurs on a semantically coherent manifold and prevents reward hacking (exploiting non-semantic noise patterns to improve the reward). Since the unconditional embedding in classifier-free guidance serves as the anchor for the model's generative distribution, Null-TTA directly steers model's generative distribution towards the target reward rather than just adjusting the samples, even without updating model parameters. Thanks to these desirable properties, we show that Null-TTA achieves state-of-the-art target test-time alignment while maintaining strong cross-reward generalisation. This establishes semantic-space optimisation as an effective and principled novel paradigm for TTA.

In this work, we explore an untapped signal in diffusion model inference. While all previous methods generate images independently at inference, we instead ask if samples can be generated collaboratively. We propose Group Diffusion, unlocking the attention mechanism to be shared across images, rather than limited to just the patches within an image. This enables images to be jointly denoised at inference time, learning both intra and inter-image correspondence. We observe a clear scaling effect -- larger group sizes yield stronger cross-sample attention and better generation quality. Furthermore, we introduce a qualitative measure to capture this behavior and show that its strength closely correlates with FID. Built on standard diffusion transformers, our GroupDiff achieves up to 32.2% FID improvement on ImageNet-256x256. Our work reveals cross-sample inference as an effective, previously unexplored mechanism for generative modeling.

We introduce MusicInfuser, an approach that aligns pre-trained text-to-video diffusion models to generate high-quality dance videos synchronized with specified music tracks. Rather than training a multimodal audio-video or audio-motion model from scratch, our method demonstrates how existing video diffusion models can be efficiently adapted to align with musical inputs. We propose a novel layer-wise adaptability criterion based on a guidance-inspired constructive influence function to select adaptable layers, significantly reducing training costs while preserving rich prior knowledge, even with limited, specialized datasets. Experiments show that MusicInfuser effectively bridges the gap between music and video, generating novel and diverse dance movements that respond dynamically to music. Furthermore, our framework generalizes well to unseen music tracks, longer video sequences, and unconventional subjects, outperforming baseline models in consistency and synchronization. All of this is achieved without requiring motion data, with training completed on a single GPU within a day.

Recent advances in foundational Video Diffusion Models (VDMs) have yielded significant progress. Yet, despite the remarkable visual quality of generated videos, reconstructing consistent 3D scenes from these outputs remains challenging, due to limited camera controllability and inconsistent generated content when viewed from distinct camera trajectories.In this paper, we propose WorldStereo, a novel framework that bridges camera-guided video generation and 3D reconstruction via two dedicated geometric memory modules. Formally, the global-geometric memory enables precise camera control while injecting coarse structural priors through incrementally updated point clouds.Moreover, the spatial-stereo memory constrains the model's attention receptive fields with 3D correspondence to focus on fine-grained details from the memory bank.These components enable WorldStereo to generate multi-view-consistent videos under precise camera control, facilitating high-quality 3D reconstruction.Furthermore, the flexible control branch-based WorldStereo shows impressive efficiency, benefiting from the distribution matching distilled VDM backbone without joint training.Extensive experiments across both camera-guided video generation and 3D reconstruction benchmarks demonstrate the effectiveness of our approach. Notably, we show that WorldStereo acts as a powerful world model, tackling diverse scene generation tasks (whether starting from perspective or panoramic images) with high-fidelity 3D results. Models will be released.

Diffusion Transformers (DiTs) achieve strong video generation performance but suffer from prohibitive computation cost due to dense spatiotemporal tokenization. Most existing works rely on uniform patchification, tokenizing non-overlapping spatiotemporal with a fixed patch size regardless of the underlying content. This content-agnostic tokenization results in substantial redundant computation, especially in visually simple or static areas. To address this inefficiency while preserving the video generation quality, we propose DynaPatch, a fine-grained dynamic patchification framework that adaptively selects patch sizes for each spatiotemporal region based on content complexity. A lightweight router predicts patch sizes directly from the latents encoded by 3D Variational Autoencoder (VAE), and is jointly optimized with the diffusion model through diffusion loss, an attention-guided saliency alignment loss, and a token-budget regularizer. Learnable patchify/unpatchify layers integrate seamlessly with standard DiT backbones, allowing flexible tokenization without architectural changes. Experiments demonstrate that DynaPatch can effectively reduce redundant computations while preserving fine details, achieving 1.3-1.8x acceleration with minimal quality degradation. On VBench, DynaPatch attains a Total Score of 83.42 at 30% token reduction, significantly outperforming prior patchification and token pruning approaches. These results indicate that content-aware patchification offers an effective direction for efficient and scalable video diffusion. Project page: https://shengli99.github.io/DynaPatch/.

The miniaturization of thermal sensors for mobile platforms inherently limits their spatial resolution and textural fidelity, leading to blurry and less informative images. Existing thermal super-resolution (SR) methods can be grouped into single-image and RGB-guided approaches: the former struggles to recover fine structures from limited information, while the latter relies on accurate and laborious cross-camera calibration, which hinders practical deployment and robustness. Here, we propose 3M-TI, a calibration-free Multi-camera cross-Modality diffusion framework for Mobile Thermal Imaging. At its core, 3M-TI integrates a cross-modal self-attention module (CSM) into the diffusion UNet, replacing the original self-attention layers to adaptively align thermal and RGB features throughout the denoising process, without requiring explicit camera calibration. This design enables the diffusion network to leverage its generative prior to enhance spatial resolution, structural fidelity, and texture detail in the super-resolved thermal images. Extensive evaluations on real-world mobile thermal cameras and public benchmarks validate our superior performance, achieving state-of-the-art results in both visual quality and quantitative metrics. More importantly, the thermal images enhanced by 3M-TI lead to substantial gains in critical downstream tasks like object detection and segmentation, underscoring its practical value for robust mobile thermal perception systems.

Video diffusion models achieve impressive visual fidelity but remain computationally prohibitive for real-time or interactive generation due to their sequential denoising process. Recent caching methods accelerate inference by reusing outputs across timesteps, typically estimating each new output from the first-order residual, which is the difference between adjacent model predictions.To mitigate the accumulated error in caching methods, we propose D2Cache, a training-free method that leverages the smoothness of second-order residual delta, which is temporal differences between consecutive first-order residuals, to predict future timesteps more accurately. We theoretically show that this second-order correction improves prediction accuracy and effectively suppresses cumulative errors. Moreover, D2Cache adaptively scales second-order deltas using error estimates derived from timestep embeddings, maintaining accuracy across varying cache intervals.Empirically, D2Cache outperforms the state-of-the-art TeaCache across four video diffusion models (Latte, Open-Sora, LTX-video, and Wan2.1) at comparable acceleration rates, showing even larger gains under higher acceleration settings.

Diffusion models have shown promising performance as data-driven priors for computational imaging, as well as some capacity to detect out-of-distribution (OOD) images. However, existing approaches to OOD detection often require some knowledge of the shifted distribution, fail to detect subtle or localized distribution shifts, and operate on full images, rather than the indirect measurements available in inverse problems. We propose an OOD detection metric based on the Kullback-Leibler divergence between the diffusion prior and the posterior distribution, that (i) does not require any calibration data or knowledge of the shifted distribution, and (ii) can detect whole images as OOD as well as localize OOD patches within an image. Experimentally, we show that this metric can detect subtle yet semantically meaningful distribution shifts, such as the shift from healthy liver CT scans to those with tumors, and generalizes across different types of diffusion models, datasets, and inverse problems.

Generating realistic conversational gestures are essential for achieving natural, socially engaging interactions with digital humans. However, existing methods typically map a single audio stream to a single speaker's motion, without considering social context or modeling the mutual dynamics between two people engaging in conversation. We present DyaDiT, a multi-modal diffusion transformer that generates contextually appropriate human motion from dyadic audio signals. Trained on Seamless Interaction Dataset, DyaDiT takes dyadic audio with optional social-context tokens to produce context-appropriate motion. It fuses information from both speakers to capture interaction dynamics, uses a motion dictionary to encode motion priors, and can optionally utilize the conversational partner's gestures to produce more responsive motion. We evaluate DyaDiT on standard motion generation metrics and conduct quantitative user studies, demonstrating that it not only surpasses existing methods on objective metrics but is also strongly preferred by users, highlighting its robustness and socially favorable motion generation. Code and models will be released upon acceptance.

While diffusion models have achieved great success in the field of video generation, this progress is accompanied by a rapidly escalating computational burden. Among the existing acceleration methods, Feature Caching is popular due to its training-free property and considerable speedup performance, but it inevitably faces semantic and detail drop with further compression. Another widely adopted method, training-aware step-distillation, though successful in image generation, also faces drastic degradation in video generation with a few steps. Furthermore, the quality loss becomes more severe when simply applying training-free feature caching to the step-distilled models, due to the sparser sampling steps. This paper novelly introduces a distillation-compatible learnable feature caching mechanism for the first time. We employ a lightweight learnable neural predictor instead of traditional training-free heuristics for diffusion models, enabling a more accurate capture of the high-dimensional feature evolution process. Furthermore, we explore the challenges of highly compressed distillation on large-scale video models and propose a conservative Restricted MeanFlow approach to achieve more stable and lossless distillation. By undertaking these initiatives, we further push the acceleration boundaries to 11.8 times while preserving generation quality. Extensive experiments demonstrate the effectiveness of our method.

We present Scalable Pixel-anchored End-to-end Diffusion (SpeeDiff), a latent diffusion method that jointly trains the VAE and the diffusion model from scratch. In principle, joint training allows the diffusion loss gradient to directly guide the VAE encoder, encouraging the formation of a generation-friendly latent space and potentially yielding faster convergence than the conventional two-stage approach with a pretrained frozen VAE. However, a naive end-to-end implementation severely degrades performance, as unrestricted backpropagation of the diffusion loss leads to latent space collapse. Our main technical contribution is a simple yet effective Tweedie Pixel Reconstruction (TPR) loss, which provides additional pixel-level feedback by decoding a predicted clean latent from an intermediate noisy state using Tweedie's formula, thereby alleviating collapse. Furthermore, our method enables jointly scaling a fully transformer-based architecture and enhances representation alignment within the end-to-end framework. Our SpeeDiff-XL model achieves over 140x and 61x faster training compared to Vanilla SiT and REPA, respectively, while attaining an FID of 1.50 without guidance on ImageNet 256x256 generation. With a more efficient 32x compressed VAE, our model further reaches an FID of 1.53 without guidance on ImageNet 512x512 generation.

Existing image SR and generic diffusion models transfer poorly to fluid SR: they are sampling-intensive, ignore physical constraints, and often yield spectral mismatch and spurious divergence. We address fluid super-resolution (SR) with **ReMD** (**Re**sidual-**M**ultigrid **D**iffusion), a physics-consistent diffusion framework. At each reverse step, ReMD performs a **multigrid residual correction**: the update direction is obtained by coupling data consistency with lightweight physics cues and then correcting the residual across scales; the multiscale hierarchy is instantiated with a **multi-wavelet** basis to capture both large structures and fine vortical details. This coarse-to-fine design accelerates convergence and preserves fine structures while remaining equation-free. Across atmospheric and oceanic benchmarks, ReMD improves accuracy and spectral fidelity, reduces divergence, and reaches comparable quality with markedly fewer sampling steps than diffusion baselines. Our results show that enforcing physics consistency **inside** the diffusion process via multigrid residual correction and multi-wavelet multiscale modeling is an effective route to efficient fluid SR.

Text-to-image (T2I) diffusion models effectively produce semantically aligned images, but their reliance on training distributions constrains their capacity for synthesizing truly novel, out-of-distribution concepts. Existing methods attempt to enhance creativity through semantic exploration, such as fusing known concept pairs, but the resulting images remain linguistically describable and confined to familiar semantic spaces. Inspired by the soft probabilistic outputs of classifiers on novel or out-of-distribution inputs, we propose Distribution-Conditional Generation, a paradigm that models novel concepts as image synthesis conditioned on class distributions, enabling controllable yet semantically unconstrained creative generation. Building on this, we propose DisTok, an encoder-decoder framework that unifies conditional and unconditional creative generation by decoding latent representations--either randomly sampled or mapped from conditions (e.g., class distributions)--into tokens representing novel concepts. DisTok is trained by iteratively sampling and fusing concept pairs from a dynamic pool to model progressively complex distributions, while enforcing semantic consistency through a vision-language model that aligns the class distributions of generated images with the input distributions. Extensive experiments demonstrate that DisTok enables efficient and flexible semantic exploration for token-level creative synthesis, achieving state-of-the-art text-image alignment and human preference.

We propose MagicQuill V2, a novel framework that introduces a layered composition paradigm to generative image editing, bridging the gap between the semantic power of modern diffusion models and the granular control of traditional graphics software. While state-of-the-art diffusion transformers excel at holistic generation, their use of singular, monolithic prompts fails to disentangle distinct user intentions for content, position, and style. To overcome this limitation, our method deconstructs creative intent into a stack of independently controllable visual cues: a content layer for what to create, a spatial layer for where to place it, a structural layer for how it is shaped, and a color layer for its palette. Our technical contributions include a specialized data generation pipeline for context-aware content integration, a unified control module to process all visual cues, and a fine-tuned spatial branch for precise local editing, including object removal. Extensive experiments validate that this layered approach effectively resolves the user intention gap, granting creators direct, intuitive, and powerful control over the generative process.