CVPR 2024 Papers

In recent advancements in high-fidelity image generation Denoising Diffusion Probabilistic Models (DDPMs) have emerged as a key player. However their application at high resolutions presents significant computational challenges. Current methods such as patchifying expedite processes in UNet and Transformer architectures but at the expense of representational capacity. Addressing this we introduce the Diffusion State Space Model (DiffuSSM) an architecture that supplants attention mechanisms with a more scalable state space model backbone. This approach effectively handles higher resolutions without resorting to global compression thus preserving detailed image representation throughout the diffusion process. Our focus on FLOP-efficient architectures in diffusion training marks a significant step forward. Comprehensive evaluations on both ImageNet and LSUN datasets at two resolutions demonstrate that DiffuSSMs are on par or even outperform existing diffusion models with attention modules in FID and Inception Score metrics while significantly reducing total FLOP usage.

Large language models (LLMs) are fine-tuned using human comparison data with Reinforcement Learning from Human Feedback (RLHF) methods to make them better aligned with users' preferences. In contrast to LLMs human preference learning has not been widely explored in text-to-image diffusion models; the best existing approach is to fine-tune a pretrained model using carefully curated high quality images and captions to improve visual appeal and text alignment. We propose Diffusion-DPO a method to align diffusion models to human preferences by directly optimizing on human comparison data. Diffusion-DPO is adapted from the recently developed Direct Preference Optimization (DPO) a simpler alternative to RLHF which directly optimizes a policy that best satisfies human preferences under a classification objective. We re-formulate DPO to account for a diffusion model notion of likelihood utilizing the evidence lower bound to derive a differentiable objective. Using the Pick-a-pic dataset of 851K crowdsourced pairwise preferences we fine-tune the base model of the state-of-the-art Stable Diffusion XL (SDXL)-1.0 model with Diffusion-DPO. Our fine-tuned base model significantly outperforms both base SDXL-1.0 and the larger SDXL-1.0 model consisting of an additional refinement model in human evaluation improving visual appeal and prompt alignment. We also develop a variant that uses AI feedback and has comparable performance to training on human preferences opening the door for scaling of diffusion model alignment methods.

Diffusion handles is a novel approach to enable 3D object edits on diffusion images requiring only existing pre-trained diffusion models depth estimation without any fine-tuning or 3D object retrieval. The edited results remain plausible photo-real and preserve object identity. Diffusion handles address a critically missing facet of generative image-based creative design. Our key insight is to lift diffusion activations for a selected object to 3D using a proxy depth 3D-transform the depth and associated activations and project them back to image space. The diffusion process guided by the manipulated activations produces plausible edited images showing complex 3D occlusion and lighting effects. We evaluate diffusion handles: quantitatively on a large synthetic data benchmark; and qualitatively by a user study showing our output to be more plausible and better than prior art at both 3D editing and identity control.

We present Diff3F as a simple robust and class-agnostic feature descriptor that can be computed for untextured input shapes (meshes or point clouds). Our method distills diffusion features from image foundational models onto input shapes. Specifically we use the input shapes to produce depth and normal maps as guidance for conditional image synthesis. In the process we produce (diffusion) features in 2D that we subsequently lift and aggregate on the original surface. Our key observation is that even if the conditional image generations obtained from multi-view rendering of the input shapes are inconsistent the associated image features are robust and hence can be directly aggregated across views. This produces semantic features on the input shapes without requiring additional data or training. We perform extensive experiments on multiple benchmarks (SHREC'19 SHREC'20 FAUST and TOSCA) and demonstrate that our features being semantic instead of geometric produce reliable correspondence across both isometric and non-isometrically related shape families. Code is available via the project webpage at https://diff3f.github.io/

Recently a number of image-mixing-based augmentation techniques have been introduced to improve the generalization of deep neural networks. In these techniques two or more randomly selected natural images are mixed together to generate an augmented image. Such methods may not only omit important portions of the input images but also introduce label ambiguities by mixing images across labels resulting in misleading supervisory signals. To address these limitations we propose DIFFUSEMIX a novel data augmentation technique that leverages a diffusion model to reshape training images supervised by our bespoke conditional prompts. First concatenation of a partial natural image and its generated counterpart is obtained which helps in avoiding the generation of unrealistic images or label ambiguities. Then to enhance resilience against adversarial attacks and improves safety measures a randomly selected structural pattern from a set of fractal images is blended into the concatenated image to form the final augmented image for training. Our empirical results on seven different datasets reveal that DIFFUSEMIX achieves superior performance compared to existing state- of-the-art methods on tasks including general classification fine-grained classification fine-tuning data scarcity and adversarial robustness.

Producing quality segmentation masks for images is a fundamental problem in computer vision. Recent research has explored large-scale supervised training to enable zero-shot transfer segmentation on virtually any image style and unsupervised training to enable segmentation without dense annotations. However constructing a model capable of segmenting anything in a zero-shot manner without any annotations is still challenging. In this paper we propose to utilize the self-attention layers in stable diffusion models to achieve this goal because the pre-trained stable diffusion model has learned inherent concepts of objects within its attention layers. Specifically we introduce a simple yet effective iterative merging process based on measuring KL divergence among attention maps to merge them into valid segmentation masks. The proposed method does not require any training or language dependency to extract quality segmentation for any images. On COCO-Stuff-27 our method surpasses the prior unsupervised zero-shot transfer SOTA method by an absolute 26% in pixel accuracy and 17% in mean IoU.

We present DiffuScene for indoor 3D scene synthesis based on a novel scene configuration denoising diffusion model. It generates 3D instance properties stored in an unordered object set and retrieves the most similar geometry for each object configuration which is characterized as a concatenation of different attributes including location size orientation semantics and geometry features. We introduce a diffusion network to synthesize a collection of 3D indoor objects by denoising a set of unordered object attributes. Unordered parametrization simplifies and eases the joint distribution approximation. The shape feature diffusion facilitates natural object placements including symmetries. Our method enables many downstream applications including scene completion scene arrangement and text-conditioned scene synthesis. Experiments on the 3D-FRONT dataset show that our method can synthesize more physically plausible and diverse indoor scenes than state-of-the-art methods. Extensive ablation studies verify the effectiveness of our design choice in scene diffusion models.

We present differentiable point-based inverse rendering DPIR an analysis-by-synthesis method that processes images captured under diverse illuminations to estimate shape and spatially-varying BRDF. To this end we adopt point-based rendering eliminating the need for multiple samplings per ray typical of volumetric rendering thus significantly enhancing the speed of inverse rendering. To realize this idea we devise a hybrid point-volumetric representation for geometry and a regularized basis-BRDF representation for reflectance. The hybrid geometric representation enables fast rendering through point-based splatting while retaining the geometric details and stability inherent to SDF-based representations. The regularized basis-BRDF mitigates the ill-posedness of inverse rendering stemming from limited light-view angular samples. We also propose an efficient shadow detection method using point-based shadow map rendering. Our extensive evaluations demonstrate that DPIR outperforms prior works in terms of reconstruction accuracy computational efficiency and memory footprint. Furthermore our explicit point-based representation and rendering enables intuitive geometry and reflectance editing.

Neural implicit surface reconstruction leveraging volume rendering has led to significant advances in multi-view reconstruction. However results for transparent objects can be very poor primarily because the rendering function fails to account for the intricate light transport induced by refraction and reflection. In this study we introduce transparent neural surface refinement (TNSR) a novel surface reconstruction framework that explicitly incorporates physical refraction and reflection tracing. Beginning with an initial approximate surface our method employs sphere tracing combined with Snell's law to cast both reflected and refracted rays. Central to our proposal is an innovative differentiable technique devised to allow signals from the photometric evidence to propagate back to the surface model by considering how the surface bends and reflects light rays. This allows us to connect surface refinement with volume rendering enabling end-to-end optimization solely on multi-view RGB images. In our experiments TNSR demonstrates significant improvements in novel view synthesis and geometry estimation of transparent objects without prior knowledge of the refractive index.

Micro-mesh (u-mesh) is a new graphics primitive for compact representation of extreme geometry consisting of a low-polygon base mesh enriched by per micro-vertex displacement. A new generation of GPUs supports this structure with hardware evolution on u-mesh ray tracing achieving real-time rendering in pixel level geometric details. In this article we present a differentiable framework to convert standard meshes into this efficient format offering a holistic scheme in contrast to the previous stage-based methods. In our construction context a u-mesh is defined where each base triangle is a parametric primitive which is then reparameterized with Laplacian operators for efficient geometry optimization. Our framework offers numerous advantages for high-quality u-mesh production: (i) end-to-end geometry optimization and displacement baking; (ii) enabling the differentiation of renderings with respect to umesh for faithful reprojectability; (iii) high scalability for integrating useful features for u-mesh production and rendering such as minimizing shell volume maintaining the isotropy of the base mesh and visual-guided adaptive level of detail. Extensive experiments on u-mesh construction for a large set of high-resolution meshes demonstrate the superior quality achieved by the proposed scheme.

In recent several years the information bottleneck (IB) principle provides an information-theoretic framework for deep multi-view clustering (MVC) by compressing multi-view observations while preserving the relevant information of multiple views. Although existing IB-based deep MVC methods have achieved huge success they rely on variational approximation and distribution assumption to estimate the lower bound of mutual information which is a notoriously hard and impractical problem in high-dimensional multi-view spaces. In this work we propose a new differentiable information bottleneck (DIB) method which provides a deterministic and analytical MVC solution by fitting the mutual information without the necessity of variational approximation. Specifically we first propose to directly fit the mutual information of high-dimensional spaces by leveraging normalized kernel Gram matrix which does not require any auxiliary neural estimator to estimate the lower bound of mutual information. Then based on the new mutual information measurement a deterministic multi-view neural network with analytical gradients is explicitly trained to parameterize IB principle which derives a deterministic compression of input variables from different views. Finally a triplet consistency discovery mechanism is devised which is capable of mining the feature consistency cluster consistency and joint consistency based on the deterministic and compact representations. Extensive experimental results show the superiority of our DIB method on 6 benchmarks compared with 13 state-of-the-art baselines.

Photometric stereo leverages variations in illumination conditions to reconstruct surface normals. Display photometric stereo which employs a conventional monitor as an illumination source has the potential to overcome limitations often encountered in bulky and difficult-to-use conventional setups. In this paper we present differentiable display photometric stereo (DDPS) addressing an often overlooked challenge in display photometric stereo: the design of display patterns. Departing from using heuristic display patterns DDPS learns the display patterns that yield accurate normal reconstruction for a target system in an end-to-end manner. To this end we propose a differentiable framework that couples basis-illumination image formation with analytic photometric-stereo reconstruction. The differentiable framework facilitates the effective learning of display patterns via auto-differentiation. Also for training supervision we propose to use 3D printing for creating a real-world training dataset enabling accurate reconstruction on the target real-world setup. Finally we exploit that conventional LCD monitors emit polarized light which allows for the optical separation of diffuse and specular reflections when combined with a polarization camera leading to accurate normal reconstruction. Extensive evaluation of DDPS shows improved normal-reconstruction accuracy compared to heuristic patterns and demonstrates compelling properties such as robustness to pattern initialization calibration errors and simplifications in image formation and reconstruction.

We describe a method for recovering the irradiance underlying a collection of images corrupted by atmospheric turbulence. Since supervised data is often technically impossible to obtain assumptions and biases have to be imposed to solve this inverse problem and we choose to model them explicitly. Rather than initializing a latent irradiance ("template") by heuristics to estimate deformation we select one of the images as a reference and model the deformation in this image by the aggregation of the optical flow from it to other images exploiting a prior imposed by Central Limit Theorem. Then with a novel flow inversion module the model registers each image TO the template but WITHOUT the template avoiding artifacts related to poor template initialization. To illustrate the robustness of the method we simply (i) select the first frame as the reference and (ii) use the simplest optical flow to estimate the warpings yet the improvement in registration is decisive in the final reconstruction as we achieve state-of-the-art performance despite its simplicity. The method establishes a strong baseline that can be further improved by integrating it seamlessly into more sophisticated pipelines or with domain-specific methods if so desired.

Audio-visual saliency prediction can draw support from diverse modality complements but further performance enhancement is still challenged by customized architectures as well as task-specific loss functions. In recent studies denoising diffusion models have shown more promising in unifying task frameworks owing to their inherent ability of generalization. Following this motivation a novel Diffusion architecture for generalized audio-visual Saliency prediction (DiffSal) is proposed in this work which formulates the prediction problem as a conditional generative task of the saliency map by utilizing input audio and video as the conditions. Based on the spatio-temporal audio-visual features an extra network Saliency-UNet is designed to perform multi-modal attention modulation for progressive refinement of the ground-truth saliency map from the noisy map. Extensive experiments demonstrate that the proposed DiffSal can achieve excellent performance across six challenging audio-visual benchmarks with an average relative improvement of 6.3% over the previous state-of-the-art results by six metrics.

We propose DiffSHEG a Diffusion-based approach for Speech-driven Holistic 3D Expression and Gesture generation. While previous works focused on co-speech gesture or expression generation individually the joint generation of synchronized expressions and gestures remains barely explored. To address this our diffusion-based co-speech motion generation Transformer enables uni-directional information flow from expression to gesture facilitating improved matching of joint expression-gesture distributions. Furthermore we introduce an outpainting-based sampling strategy for arbitrary long sequence generation in diffusion models offering flexibility and computational efficiency. Our method provides a practical solution that produces high-quality synchronized expression and gesture generation driven by speech. Evaluated on two public datasets our approach achieves state-of-the-art performance both quantitatively and qualitatively. Additionally a user study confirms the superiority of our method over prior approaches. By enabling the real-time generation of expressive and synchronized motions our method showcases its potential for various applications in the development of digital humans and embodied agents.

This paper endeavors to advance the precision of snapshot compressive imaging (SCI) reconstruction for multispectral image (MSI). To achieve this we integrate the advantageous attributes of established SCI techniques and an image generative model propose a novel structured zero-shot diffusion model dubbed DiffSCI. DiffSCI leverages the structural insights from the deep prior and optimization-based methodologies complemented by the generative capabilities offered by the contemporary denoising diffusion model. Specifically firstly we employ a pre-trained diffusion model which has been trained on a substantial corpus of RGB images as the generative denoiser within the Plug-and-Play framework for the first time. This integration allows for the successful completion of SCI reconstruction especially in the case that current methods struggle to address effectively. Secondly we systematically account for spectral band correlations and introduce a robust methodology to mitigate wavelength mismatch thus enabling seamless adaptation of the RGB diffusion model to MSIs.Thirdly an accelerated algorithm is implemented to expedite the resolution of the data subproblem. This augmentation not only accelerates the convergence rate but also elevates the quality of the reconstruction process.We present extensive testing to show that DiffSCI exhibits discernible performance enhancements over prevailing self-supervised and zero-shot approaches surpassing even supervised transformer counterparts across both simulated and real datasets. Code is at https://github.com/PAN083/DiffSCI.

We present DiffPortrait3D a conditional diffusion model that is capable of synthesizing 3D-consistent photo-realistic novel views from as few as a single in-the-wild portrait. Specifically given a single RGB input we aim to synthesize plausible but consistent facial details rendered from novel camera views with retained both identity and facial expression. In lieu of time-consuming optimization and fine-tuning our zero-shot method generalizes well to arbitrary face portraits with unposed camera views extreme facial expressions and diverse artistic depictions. At its core we leverage the generative prior of 2D diffusion models pre-trained on large-scale image datasets as our rendering backbone while the denoising is guided with disentangled attentive control of appearance and camera pose. To achieve this we first inject the appearance context from the reference image into the self-attention layers of the frozen UNets. The rendering view is then manipulated with a novel conditional control module that interprets the camera pose by watching a condition image of a crossed subject from the same view. Furthermore we insert a trainable cross-view attention module to enhance view consistency which is further strengthened with a novel 3D-aware noise generation process during inference. We demonstrate state-of-the-art results both qualitatively and quantitatively on our challenging in-the-wild and multi-view benchmarks.

Existing diffusion models for pose-guided human video generation mostly suffer from temporal inconsistency in the generated appearance and poses due to the inherent randomization nature of the generation process. In this paper we propose a novel framework DiffPerformer to synthesize high-fidelity and temporally consistent human video. Without complex architecture modification or costly training DiffPerformer finetunes a pretrained diffusion model on a single video of the target character and introduces an implicit video representation as a proxy to learn temporally consistent guidance for the diffusion model. The guidance is encoded into VAE latent space and an iterative optimization loop is constructed between the implicit video representation and the diffusion model allowing to harness the smooth property of the implicit video representation and the generative capabilities of the diffusion model in a mutually beneficial way. Moreover we propose 3D-aware human flow as a temporal constraint during the optimization to explicitly model the correspondence between driving poses and human appearance. This alleviates the misalignment between guided poses and target performer and therefore maintains the appearance coherence under various motions. Extensive experiments demonstrate that our method outperforms the state-of-the-art methods.