3D Gaussian Splatting (3DGS) has begun incorporating rich information from 2D foundation models. However, most approaches rely on a bottom-up optimization process that treats raw 2D features as ground truth, incurring increased computational costs. We propose a top-down pipeline for constructing compact and fast 3D feature fields, namely, CF3. We first perform a fast weighted fusion of multi-view 2D features with pre-trained Gaussians. This approach enables training a per-Gaussian autoencoder directly on the lifted features, instead of training autoencoders in the 2D domain. As a result, the autoencoder better aligns with the inference-time feature distribution, achieving high compression fidelity and storage efficiency. Finally, we introduce an adaptive sparsification method that prunes and merges Gaussians, constructing an efficient representation while maintaining details. Our approach achieves a competitive 3D feature field while using as little as 5% of the Gaussians compared to Feature-3DGS.
论文检索
输入标题、作者或关键词，从 1,092 篇学术成果中精准定位
Bridging different modalities lies at the heart of cross-modality generation. While conventional approaches treat the text modality as a conditioning signal that gradually guides the denoising process from Gaussian noise to the target image modality, we explore a much simpler paradigm--directly evolving between text and image modalities through flow matching. This requires projecting both modalities into a shared latent space, which poses a significant challenge due to their inherently different representations: text is highly semantic and encoded as 1D tokens, whereas images are spatially redundant and represented as 2D latent embeddings. To address this, we introduce FlowTok, a minimal framework that seamlessly flows across text and images by encoding images into a compact 1D token representation. Compared to prior methods, this design reduces the latent space size by 3.3x at an image resolution of 256, eliminating the need for complex conditioning mechanisms or noise scheduling. Moreover, FlowTok naturally extends to image-to-text generation under the same formulation. With its streamlined architecture centered around compact 1D tokens, FlowTok is highly memory-efficient, requires significantly fewer training resources, and achieves much faster sampling speeds--all while delivering performance comparable to state-of-the-art models. Code is available at https://github.com/TACJu/FlowTok.
GaussianFlowOcc: Sparse and Weakly Supervised Occupancy Estimation using Gaussian Splatting and Temporal Flow
PDF ↗Occupancy estimation has become a prominent task in 3D computer vision, particularly within the autonomous driving community. In this paper, we present a novel approach to occupancy estimation, termed GaussianFlowOcc, which is inspired by Gaussian Splatting and replaces traditional dense voxel grids with a sparse 3D Gaussian representation. Our efficient model architecture based on a Gaussian Transformer significantly reduces computational and memory requirements by eliminating the need for expensive 3D convolutions used with inefficient voxel-based representations that predominantly represent empty 3D spaces. GaussianFlowOcc effectively captures scene dynamics by estimating temporal flow for each Gaussian during the overall network training process, offering a straightforward solution to a complex problem that is often neglected by existing methods. Moreover, GaussianFlowOcc is designed for scalability, as it employs weak supervision and does not require costly dense 3D voxel annotations based on additional data (e.g., LiDAR). Through extensive experimentation, we demonstrate that GaussianFlowOcc significantly outperforms all previous methods for weakly supervised occupancy estimation on the nuScenes dataset while featuring an inference speed that is 50 times faster than current SOTA.
Explicitly disentangling style and content in vision models remains challenging due to their semantic overlap and the subjectivity of human perception. Existing methods propose separation through generative or discriminative objectives, but they still face the inherent ambiguity of disentangling intertwined concepts. Instead, we ask: Can we bypass explicit disentanglement by learning to merge style and content invertibly, allowing separation to emerge naturally? We propose SCFlow, a flow-matching framework that learns bidirectional mappings between entangled and disentangled representations. Our approach is built upon three key insights: 1) Training solely to merge style and content, a well-defined task, enables invertible disentanglement without explicit supervision; 2) flow matching bridges on arbitrary distributions, avoiding the restrictive Gaussian priors of diffusion models and normalizing flows; and 3) a synthetic dataset of 510,000 samples was curated to simulate disentanglement through systematic style-content pairing. Beyond controllable generation tasks, we demonstrate that SCFlow generalizes to ImageNet-1k and WikiArt in zero-shot settings and achieves competitive performance, highlighting that disentanglement naturally emerges from the invertible merging process. Code and dataset: https://github.com/CompVis/SCFlow
Recent works in 3D representation learning and multimodal pre-training have made remarkable progress. However, typically multimodal 3D models are only capable of handling point clouds. Compared to the emerging 3D representation technique, 3D Gaussian Splatting (3DGS), the spatially sparse point cloud cannot depict the texture information of 3D objects, resulting in inferior reconstruction capabilities. This limitation constrains the potential of point cloud-based 3D multimodal representation learning. In this paper, we present CLIP-GS, a novel multimodal representation learning framework grounded in 3DGS. We introduce the GS Tokenizer to generate serialized gaussian tokens, which are then processed through a series of transformer layers pre-initialized with weights from point cloud models, resulting in the 3DGS embeddings. CLIP-GS leverages contrastive loss between 3DGS and the visual-text embeddings of CLIP, and we introduce an image voting loss to guide the directionality and convergence of gradient optimization. Furthermore, we develop an efficient way to generate triplets of 3DGS, images, and text, facilitating CLIP-GS in learning unified multimodal representations. Leveraging the well-aligned multimodal representations, CLIP-GS demonstrates versatility and outperforms point cloud-based models on various 3D tasks, including multimodal retrieval, zero-shot, and few-shot classification.
In recent years, neural networks have achieved significant progress in offline image processing. However, in online scenarios, particularly in on-chip implementations, memory usage emerges as a critical bottleneck due to the limited memory resources of integrated image processors. In this study, we focus on reducing the memory footprint of neural networks for on-chip image processing by optimizing network design for efficient memory utilization. Specifically, we consider a typical scenario in which images outputted from an image sensor are processed sequentially using line buffers in a line-by-line manner. This setting necessitates the modeling of both intra-line and inter-line correlations--capturing dependencies among pixels within a single line group and across different line groups, respectively.To model intra-line correlations, we propose a progressive feature enhancement strategy, where line pixels are processed with expanding strip convolutions in multiple stages.For inter-line correlation modeling, we introduce a hierarchical line buffer formulation, where features extracted from previous lines are incrementally reused and compressed across multiple hierarchical levels.Comprehensive experiments on various image processing tasks, including RAW denoising, Gaussian denoising, and super-resolution, demonstrate that the proposed method achieves a superior trade-off between performance and memory efficiency than previous solutions, e.g., up to 1dB PSNR gain in RAW denoising at one-fifth of peak memory usage.
3D generation has made significant progress, however, it still largely remains at the object-level. Feedforward 3D scene-level generation has been rarely explored due to the lack of models capable of scaling-up latent representation learning on 3D scene-level data. Unlike object-level generative models, which are trained on well-labeled 3D data in a bounded canonical space, scene-level generations with 3D scenes represented by 3D Gaussian Splatting (3DGS) are unbounded and exhibit scale inconsistency across different scenes, making unified latent representation learning for generative purposes extremely challenging. In this paper, we introduce Can3Tok, the first 3D scene-level variational autoencoder (VAE) capable of encoding a large number of Gaussian primitives into a low-dimensional latent embedding, which effectively captures both semantic and spatial information of the inputs. Beyond model design, we propose a general pipeline for 3D scene data processing to address scale inconsistency issue. We validate our method on the recent scene-level 3D dataset DL3DV-10K, where we found that only Can3Tok successfully generalizes to novel 3D scenes, while compared methods fail to converge on even a few hundred scene inputs during training and exhibit zero generalization ability during inference. Finally, we demonstrate image-to-3DGS and text-to-3DGS generation as our applications to demonstrate it's ability to faciliate downstream generation tasks. Code will be released.
Video large language models (Video-LLMs) have made significant progress in understanding videos. However, processing multiple frames leads to lengthy visual token sequences, presenting challenges such as the limited context length cannot accommodate the entire video, and the inclusion of irrelevant frames hinders visual perception. Hence, effective frame selection is crucial. This paper emphasizes that frame selection should follow three key principles: query relevance, list-wise diversity, and sequentiality. Existing methods, such as uniform frame sampling and query-frame matching, do not capture all of these principles. Thus, we propose Markov decision determinantal point process with dynamic programming (MDP^3) for frame selection, a training-free and model-agnostic method that can be seamlessly integrated into existing Video-LLMs. Our method first estimates frame similarities conditioned on the query using a conditional Gaussian kernel within the reproducing kernel Hilbert space (RKHS). We then apply the determinantal point process (DPP) to the similarity matrix to capture both query relevance and list-wise diversity. To incorporate sequentiality, we segment the video and apply DPP within each segment, conditioned on the preceding segment selection, modeled as a Markov decision process (MDP) for allocating selection sizes across segments. Theoretically, MDP^3 provides a (1-1/e)-approximate solution to the NP-hard list-wise frame selection problem with pseudo-polynomial time complexity, demonstrating its efficiency. Empirically, MDP significantly outperforms existing methods, verifying its effectiveness and robustness.
Gaussian-based World Model: Gaussian Priors for Voxel-Based Occupancy Prediction and Future Motion Prediction
PDF ↗In autonomous driving, accurately predicting occupancy and motion is crucial for safe navigation within dynamic environments. However, existing methods often suffer from difficulties in handling complex scenes and uncertainty arising from sensor data. To address these issues, we propose a new Gaussian-based World Model (GWM), seamlessly integrating raw multi-modal sensor inputs. In 1st stage, Gaussian representation learner utilizes self-supervised pretraining to learn robust Gaussian representation. Gaussian representation integrates semantic and geometric information and establishes a robust probabilistic understanding of the environment. In 2nd stage, GWM seamlessly integrates learning, simulation, and planning into a unified framework, empowering the uncertainty-aware simulator & planner to jointly forecast future scene evolutions and vehicle trajectories. Simulator generates future scene predictions by modeling both static and dynamic elements, while planner calculates optimal paths to minimize collision risks, thus enhancing navigation safety. Overall, GWM employs a sensor-to-planning world model that directly processes raw sensor data, setting it apart from previous methods. Experiments show that GWM outperforms state-of-the-art approaches by 1.46% in semantic comprehension and 0.07m in motion prediction. Moreover, we provide an in-depth analysis of Gaussian representations under complex scenarios.
DynFaceRestore: Balancing Fidelity and Quality in Diffusion-Guided Blind Face Restoration with Dynamic Blur-Level Mapping and Guidance
PDF ↗Blind Face Restoration aims to recover high-fidelity, detail-rich facial images from unknown degraded inputs, presenting significant challenges in preserving both identity and detail. Pre-trained diffusion models have been increasingly used as image priors to generate fine details. Still, existing methods often use fixed diffusion sampling timesteps and a global guidance scale, assuming uniform degradation. This limitation and potentially imperfect degradation kernel estimation frequently lead to under- or over-diffusion, resulting in an imbalance between fidelity and quality. We propose DynFaceRestore, a novel blind face restoration approach that learns to map any blindly degraded input to Gaussian blurry images. By leveraging these blurry images and their respective Gaussian kernels, we dynamically select the starting timesteps for each blurry image and apply closed-form guidance during the diffusion sampling process to maintain fidelity. Additionally, we introduce a dynamic guidance scaling adjuster that modulates the guidance strength across local regions, enhancing detail generation in complex areas while preserving structural fidelity in contours. This strategy effectively balances the trade-off between fidelity and quality. DynFaceRestore achieves state-of-the-art performance in both quantitative and qualitative evaluations, demonstrating robustness and effectiveness in blind face restoration.
We introduce HouseTour, a method for spatially-aware 3D camera trajectory and natural language summary generation from a collection of images depicting an existing 3D space. Unlike existing vision-language models (VLMs), which struggle with geometric reasoning, our approach generates smooth video trajectories via a diffusion process constrained by known camera poses and integrates this information into the VLM for 3D-grounded descriptions. We synthesize the final video using 3D Gaussian splatting to render novel views along the trajectory. To support this task, we present the HouseTour dataset, which includes over 1,200 house-tour videos with camera poses, 3D reconstructions, and real estate descriptions. Experiments demonstrate that incorporating 3D camera trajectories into the text generation process improves performance over methods handling each task independently. We evaluate both individual and end-to-end performance, introducing a new joint metric. Our work enables automated, professional-quality video creation for real estate and touristic applications without requiring specialized expertise or equipment.
3D Gaussian Splatting (3DGS) has demonstrated its potential in reconstructing scenes from unposed images. However, optimization-based 3DGS methods struggle with sparse views due to limited prior knowledge. Meanwhile, feed-forward Gaussian approaches are constrained by input formats, making it challenging to incorporate more input views. To address these challenges, we propose RegGS, a 3D Gaussian registration-based framework for reconstructing unposed sparse views. RegGS aligns local 3D Gaussians generated by a feed-forward network into a globally consistent 3D Gaussian representation. Technically, we implement an entropy-regularized Sinkhorn algorithm to efficiently solve the optimal transport Mixture 2-Wasserstein MW_2 distance, which serves as an alignment metric for Gaussian mixture models (GMMs) in Sim(3) space. Furthermore, we design a joint 3DGS registration module that integrates the MW_2 distance, photometric consistency, and depth geometry. This enables a coarse-to-fine registration process while accurately estimating camera poses and aligning the scene. Experiments on the RE10K and ACID datasets demonstrate that RegGS effectively registers local Gaussians with high fidelity, achieving precise pose estimation and high-quality novel view synthesis.
ContraGS: Codebook-Condensed and Trainable Gaussian Splatting for Fast, Memory-Efficient Reconstruction
PDF ↗3D Gaussian Splatting (3DGS) is a state-of-art technique to model real-world scenes with high quality and real-time rendering.Typically, a higher quality representation can be achieved by using a large number of 3D Gaussians. However, using large 3D Gaussian counts significantly increases the GPU device memory for storing model parameters. A large model thus requires powerful GPUs with high memory capacities for training and has slower training/rendering latencies due to the inefficiencies of memory access and data movement. In this work, we introduce ContraGS, a method to enable training directly on compressed 3DGS representations without reducing the Gaussian Counts, and thus with a little loss in model quality. ContraGS leverages codebooks to compactly store a set of Gaussian parameter vectors throughout the training process, thereby significantly reducing memory consumption. While codebooks have been demonstrated to be highly effective at compressing fully trained 3DGS models, directly training using codebook representations is an unsolved challenge. ContraGS solves the problem of learning non-differentiable parameters in codebook-compressed representations by posing parameter estimation as a Bayesian inference problem. To this end, ContraGS provides a framework that effectively uses MCMC sampling to sample over a posterior distribution of these compressed representations. With ContraGS, we demonstrate that ContraGS significantly reduces the peak memory during training (on average 3.49X) and accelerated training and rendering 1.36Xand 1.88X on average, respectively), while retraining close to state-of-art quality.
The ability to abstract complex 3D environments into simplified and structured representations is crucial across various domains. 3D semantic scene graphs (SSGs) achieve this by representing objects as nodes and their interrelationships as edges, facilitating high-level scene understanding. Existing methods for 3D SSG generation, however, face significant challenges, including high computational demands and non-incremental processing that hinder their suitability for real-time open-world applications. To address this issue, we propose FROSS (Faster-than-Real-Time Online 3D Semantic Scene Graph Generation), an innovative approach for online and faster-than-real-time 3D SSG generation that leverages the direct lifting of 2D scene graphs to 3D space and represents objects as 3D Gaussian distributions. This framework eliminates the dependency on precise and computationally-intensive point cloud processing. Furthermore, we extend the Replica dataset with inter-object relationship annotations, creating the ReplicaSSG dataset for comprehensive evaluation of FROSS. The experimental results from evaluations on ReplicaSSG and 3DSSG datasets show that FROSS can achieve superior performance while operating significantly faster than prior 3D SSG generation methods. Our implementation and dataset are publicly available at https://github.com/Howardkhh/FROSS.
ReconDreamer++: Harmonizing Generative and Reconstructive Models for Driving Scene Representation
PDF ↗Combining reconstruction models with generative models has emerged as a promising paradigm for closed-loop simulation in autonomous driving. For example, ReconDreamer has demonstrated remarkable success in rendering large-scale maneuvers. However, a significant gap remains between the generated data and real-world sensor observations, particularly in terms of fidelity for structured elements, such as the ground surface. To address these challenges, we propose ReconDreamer++, an enhanced framework that significantly improves the overall rendering quality by mitigating the domain gap and refining the representation of the ground surface.Specifically, ReconDreamer++ introduces the Novel Trajectory Deformable Network (NTDNet), which leverages learnable spatial deformation mechanisms to bridge the domain gap between synthesized novel views and original sensor observations. Moreover, for structured elements such as the ground surface, we preserve geometric prior knowledge in 3D Gaussians, andthe optimization process focuses on refining appearance attributes while preserving the underlying geometric structure. Experimental evaluations conducted on multiple datasets (Waymo, nuScenes, PandaSet, and EUVS) confirm the superior performance of ReconDreamer++. Specifically, on Waymo, ReconDreamer++ achieves performance comparable to Street Gaussians for the original trajectory while significantly outperforming ReconDreamer on novel trajectories. In particular, it achieves substantial improvements, including a 6.1% increase in NTA-IoU, a 23. 0% improvement in FID, and a remarkable 4.5% gain in the ground surface metric NTL-IoU, highlighting its effectiveness in accurately reconstructing structured elements such as the road surface.
The evolving media landscape increasingly demands immersive, non-linear formats supported by innovative tools for content creation and distribution. The Horizon Europe XReco project addresses this need by providing a unified, data-driven ecosystem for next-generation media production, with a focus on extended reality (XR) and virtual production. The platform integrates ingestion of diverse media types (text, images, audio, video, 3D), cross-modal search, 3D content creation, sharing, and monetization. Central to XReco is a metadata-driven ingestion system that overcomes archive fragmentation by enabling efficient organization and access to content from sources like broadcasters, online news, and open repositories. This capability was demonstrated through a short TV documentary on Guglielmo Marconi, created using historical materials assembled via the XReco platform. The platform's Orchestrator module empowers users with powerful cross-modal semantic search capabilities, leveraging neural descriptors to enable queries across different media formats. Editorial teams can retrieve relevant contents searching by keywords like ''telegraph'' or perform reverse image searches to identify and contextualize visual assets like images and 3D models. This unified search functionality significantly enhances content discovery and reuse. A major innovation of the platform consists in providing a set of tools for enhancing the quality of the ingested contents, as well as generating 3D models from 2D assets using state-of-the-art techniques (video super resolution, blind face restoration, NeRF, Gaussian Splatting, Structure from Motion). These services are accessible and tunable via a unified interface, which provides a streamlined user experience and hides the complexity of the underlying technologies. For the Marconi documentary, detailed 3D models of key technological artifacts were created, enabling viewers to interactively explore these objects from multiple perspectives. XReco also supports seamless integration with third-party tools to enrich production workflows. Our documentary incorporated photorealistic digital avatars created with Unreal MetaHuman, animated via motion capture, and featured holoported human experts alongside real presenters within dynamic virtual environments. A noteworthy example is the virtual reconstruction of the RAI Radio Museum in Turin based on Gaussian Splatting, in which avatars from remote locations are developed with Unity and rendered using 4D Gaussian Splatting and Free Viewpoint Video (FVV) technologies. Compatibility with platforms such as Unity and Unreal Engine further facilitates the creation of visually compelling XR experiences. In summary, the XReco platform represents a robust end-to-end solution that effectively tackles the technical and commercial complexities of modern XR and virtual production, paving the way for innovative storytelling in the evolving media ecosystem. During the demo, attendees will have a walkthrough of the platform functionalities, highlighting key technologies for content search, filtering, and processing. They will also be able to enjoy a short documentary about Guglielmo Marconi, produced by our editorial team using XReco technology. After the walkthrough, attendees will have the opportunity to interact directly with the XReco platform to explore its features hands-on-such as testing the search capabilities, creating 3D assets, and experimenting with other available tools. This will provide a more engaging and comprehensive experience of the demo's functionalities. Link to the video: https://drive.google.com/drive/folders/15XTkg-x1U62hQ2dRo2ABiT3LG94CJXO5
Point cloud processing and 3D vision have emerged as very hot topics in the multimedia community. Point clouds can give an immersive visual experience and provide accurate structural information of 3D objects and scenes in the applications including virtual reality/augmented reality (VR/AR), autonomous driving, robot navigation, and geo-information systems (GIS). Moreover, 3D Gaussian splatting has become a very powerful and popular tool for 3D reconstruction, generation and rendering, as well as compression, representation and understanding. 3D Gaussian splatting technology can also be deemed as an extension of point cloud processing technology. 3D vision technologies empower the developments of immersive media, embodied artificial intelligence (AI) and unmanned systems. Their challenges in processing, analysis, and applications have attracted significant interest from industry, academia, and standardization bodies. This workshop invites innovative contributions in point cloud processing and 3D Gaussian splatting to propel the advancements of 3D vision technologies.
3D Gaussian Splatting (3DGS) is an emerging media representation that reconstructs real-world 3D scenes in high fidelity, enabling 6-degrees-of-freedom (6-DoF) navigation in virtual reality (VR). However, developing and evaluating 3DGS-enabled applications and optimizing their rendering performance require realistic user navigation data. Such data is currently unavailable for photorealistic 3DGS reconstructions of real-world scenes. This paper introduces EyeNavGS, the first publicly available 6-DoF navigation dataset featuring traces from 46 participants exploring twelve diverse, real-world 3DGS scenes. The dataset was collected at two sites, using the Meta Quest Pro headsets, recording the head pose and eye gaze data for each rendered frame during free world standing 6-DoF navigation. For each of the twelve scenes, we performed careful scene initialization to correct for scene tilt and scale, ensuring a perceptually-comfortable VR experience. We also release our open-source SIBR viewer software fork with record-and-replay functionalities and a suite of utility tools for data processing, conversion, and visualization. The EyeNavGS dataset and its accompanying software tools provide valuable resources for advancing research in 6-DoF viewport prediction, adaptive streaming, 3D saliency, and foveated rendering for 3DGS scenes. The EyeNavGS dataset is available at: https://symmru.github.io/EyeNavGS/
The growing popularity of 3D Gaussian Splatting (3DGS) has intensified the need for effective copyright protection. Current 3DGS watermarking methods rely on computationally expensive fine-tuning procedures for each predefined message. We propose the first generalizable watermarking framework that enables efficient protection of Splatter Image-based 3DGS models through a single forward pass. We introduce GaussianBridge that transforms unstructured 3D Gaussians into Splatter Image format, enabling direct neural processing for arbitrary message embedding. To ensure imperceptibility, we design a Gaussian-Uncertainty-Perceptual heatmap prediction strategy for preserving visual quality. For robust message recovery, we develop a dense segmentation-based extraction mechanism that maintains reliable extraction even when watermarked objects occupy minimal regions in rendered views. Project page: https://kevinhuangxf.github.io/marksplatter.
3D Gaussian Splatting (3DGS) is a recent popular technique that can reconstruct the radiance field representation of the scene efficiently. However, the naive 3DGS algorithm is easily affected by noisy pixels from transient and dynamic objects. To resolve this matter and enable the robust learning of 3DGS, previous work proposed to generate a binary mask from per-pixel training loss or an image segmentation result. However, such modeling of the distractors is not adaptive to the 3DGS model learning process, and might lead to wrong identification of noisy pixels and would affect the reconstruction performance. Instead, we propose to learn a soft mask of the likelihood of the distractors. Moreover, we develop techniques to model the spatial pattern of distractors and learn them from a design curriculum, avoiding confusion between clean and noisy pixels. Our method demonstrates state-of-the-art performance on robust novel view synthesis from distractor images, evaluated on major benchmarks of this task.