In this paper, we present KP-RED, a unified KeyPoint-driven REtrieval and Deformation framework that takes object scans as input and jointly retrieves and deforms the most geometrically similar CAD models from a pre-processed database to tightly match the target. Unlike existing dense matching based methods that typically struggle with noisy partial scans, we propose to leverage category-consistent sparse keypoints to naturally handle both full and partial object scans. Specifically, we first employ a lightweight retrieval module to establish a keypoint-based embedding space, measuring the similarity among objects by dynamically aggregating deformation-aware local-global features around extracted keypoints. Objects that are close in the embedding space are considered similar in geometry. Then we introduce the neural cage-based deformation module that estimates the influence vector of each keypoint upon cage vertices inside its local support region to control the deformation of the retrieved shape. Extensive experiments on the synthetic dataset PartNet and the real-world dataset Scan2CAD demonstrate that KP-RED surpasses existing state-of-the-art approaches by a large margin. Codes and trained models will be released in https://github.com/lolrudy/KP-RED.
During the Gaussian Splatting optimization process, the scene's geometry can gradually deteriorate if its structure is not deliberately preserved, especially in non-textured regions such as walls, ceilings, and furniture surfaces. This degradation significantly affects the rendering quality of novel views that deviate significantly from the viewpoints in the training data. To mitigate this issue, we propose a novel approach called GeoGaussian. Based on the smoothly connected areas observed from point clouds, this method introduces a novel pipeline to initialize thin Gaussians aligned with the surfaces, where the characteristic can be transferred to new generations through a carefully designed densification strategy. Finally, the pipeline ensures that the scene's geometry and texture are maintained through constrained optimization processes with explicit geometry constraints. Benefiting from the proposed architecture, the generative ability of 3D Gaussians is enhanced, especially in structured regions. Our proposed pipeline achieves state-of-the-art performance in novel view synthesis and geometric reconstruction, as evaluated qualitatively and quantitatively on public datasets.
Precisely perceiving the geometric and semantic properties of real-world 3D objects is crucial for the continued evolution of augmented reality and robotic applications. To this end, we present \algfull{} (\algname{}), which incorporates vision-language embeddings of foundation models into 3D Gaussian Splatting (GS). The key contribution of this work is an efficient method to reconstruct and represent 3D vision-language models. This is achieved by distilling feature maps generated from image-based foundation models into those rendered from our 3D model. To ensure high-quality rendering and fast training, we introduce a novel scene representation by integrating strengths from both GS and multi-resolution hash encodings (MHE). Our effective training procedure also introduces a pixel alignment loss that makes the rendered feature distance of same semantic entities close, following the pixel-level semantic boundaries. Our results demonstrate remarkable multi-view semantic consistency, facilitating diverse downstream tasks, beating state-of-the-art methods by $\mathbf{10.2}$ percent on open-vocabulary language-based object detection, despite that we are $\mathbf{851\times}$ faster for inference. This research explores the intersection of vision, language, and 3D scene representation, paving the way for enhanced scene understanding in uncontrolled real-world environments. We plan to release the code upon paper acceptance.
Reconstructing hand-held objects from a single RGB image without known 3D object templates, category prior, or depth information is a vital yet challenging problem in computer vision. In contrast to prior works that utilize deterministic modeling paradigms, which make it hard to account for the uncertainties introduced by hand- and self-occlusion, we employ a probabilistic point cloud denoising diffusion model to tackle the above challenge. In this work, we present Hand-Aware Conditional Diffusion for monocular hand-held object reconstruction (HACD), modeling the hand-object interaction in two aspects. First, we introduce hand-aware conditioning to model hand-object interaction from both semantic and geometric perspectives. Specifically, a unified hand-object semantic embedding compensates for the 2D local feature deficiency induced by hand occlusion, and a hand articulation embedding further encodes the relationship between object vertices and hand joints. Second, we propose a hand-constrained centroid fixing scheme, which utilizes hand vertices priors to restrict the centroid deviation of partially denoised point cloud during diffusion and reverse process. Removing the centroid bias interference allows the diffusion models to focus on the reconstruction of shape, thus enhancing the stability and precision of local feature projection. Experiments on the synthetic ObMan dataset and two real-world datasets, HO3D and MOW, demonstrate our approach surpasses all existing methods by a large margin.
In this work, we present a novel dense-correspondence method for 6DoF object pose estimation from a single RGB-D image. While many existing data-driven methods achieve impressive performance, they tend to be time-consuming due to their reliance on rendering-based refinement approaches. To circumvent this limitation, we present HiPose, which establishes 3D-3D correspondences in a coarse-to-fine manner with a hierarchical binary surface encoding. Unlike previous dense-correspondence methods, we estimate the correspondence surface by employing point-to-surface matching and iteratively constricting the surface until it becomes a correspondence point while gradually removing outliers. Extensive experiments on public benchmarks LM-O, YCB-V, and T-Less demonstrate that our method surpasses all refinement-free methods and is even on par with expensive refinement-based approaches. Crucially, our approach is computationally efficient and enables real-time critical applications with high accuracy requirements. Code and models will be released.
Category-level object pose estimation, aiming to predict the 6D pose and 3D size of objects from known categories, typically struggles with large intra-class shape variation. Existing works utilizing mean shapes often fall short of capturing this variation. To address this issue, we present SecondPose, a novel approach integrating object-specific geometric features with semantic category priors from DINOv2. Leveraging the advantage of DINOv2 in providing SE(3)-consistent semantic features, we hierarchically extract two types of SE(3)-invariant geometric features to further encapsulate local-to-global object-specific information. These geometric features are then point-aligned with DINOv2 features to establish a consistent object representation under SE(3) transformations, facilitating the mapping from camera space to the pre-defined canonical space, thus further enhancing pose estimation. Extensive experiments on NOCS-REAL275 demonstrate that SecondPose achieves a 12.4% leap forward over the state-of-the-art. Moreover, on a more complex dataset HouseCat6D which provides photometrically challenging objects, SecondPose still surpasses other competitors by a large margin. The code will be released soon.
In this paper, we present ShapeMaker, a unified self-supervised learning framework for joint shape canonicalization, segmentation, retrieval and deformation. Given a partially-observed object in an arbitrary pose, we first canonicalize the object by extracting point-wise affine-invariant features, disentangling inherent structure of the object with its pose and size. These learned features are then leveraged to predict semantically consistent part segmentation and corresponding part centers. Next, our lightweight retrieval module aggregates the features within each part as its retrieval token and compare all the tokens with source shapes from a pre-established database to identify the most geometrically similar shape. Finally, we deform the retrieved shape in the deformation module to tightly fit the input object by harnessing part center guided neural cage deformation. The key insight of ShapeMaker is the simultaneous training of the four highly-associated processes: canonicalization, segmentation, retrieval, and deformation, leveraging cross-task consistency losses for mutual supervision. Extensive experiments on synthetic datasets PartNet, ComplementMe, and real-world dataset Scan2CAD demonstrate that ShapeMaker surpasses competitors by a large margin. Codes will be released soon.
Comprehending natural language instructions is a critical skill for robots to cooperate effectively with humans. In this paper, we aim to learn 6D poses for roboticassembly by natural language instructions. For this purpose, Language-Instructed 6D Pose Regression Network (LanPose) is proposed to jointly predict the 6D poses of the observed object and the corresponding assembly position. Our proposed approach is based on the fusion of geometric and linguistic features, which allows us to finely integrate multi-modality input and map it to the 6D pose in SE(3) space by the cross-attention mechanism and the language-integrated 6D pose mapping module, respectively. To validate the effectiveness of our approach, an integrated robotic system is established to precisely and robustly perceive, grasp, manipulate and assemble blocks by language commands. 98.09 and 93.55 in ADD(-S)-0.1d are derived for the prediction of 6D object pose and 6D assembly pose, respectively. Both quantitative and qualitative results demonstrate the effectiveness of our proposed language-instructed 6D pose estimation methodology and its potential to enable robots to better understand and execute natural language instructions.
Previous works concerning single-view hand-held object reconstruction typically utilize supervision from 3D ground truth models, which are hard to collect in real world. In contrast, abundant videos depicting hand-object interactions can be accessed easily with low cost, although they only give partial object observations with complex occlusion. In this paper, we present MOHO to reconstruct hand-held object from a single image with multi-view supervision from hand-object videos, tackling two predominant challenges including object's self-occlusion and hand-induced occlusion. MOHO inputs semantic features indicating visible object parts and geometric embeddings provided by hand articulations as partial-to-full cues to resist object's self-occlusion, so as to recover full shape of the object. Meanwhile, a novel 2D-3D hand-occlusion-aware training scheme following the synthetic-to-real paradigm is proposed to release hand-induced occlusion. In the synthetic pre-training stage, 2D-3D hand-object correlations are constructed by supervising MOHO with rendered images to complete the hand-concealed regions of the object in both 2D and 3D space. Subsequently, MOHO is finetuned in real world by the mask-weighted volume rendering supervision adopting hand-object correlations obtained during pre-training. Extensive experiments on HO3D and DexYCB datasets demonstrate that 2D-supervised MOHO gains superior results against 3D-supervised methods by a large margin. Codes and key assets will be released soon.
Object rearrangement is pivotal in robotic-environment interactions, representing a significant capability in embodied AI. In this paper, we present SG-Bot, a novel rearrangement framework that utilizes a coarse-to-fine scheme with a scene graph as the scene representation. Unlike previous methods that rely on either known goal priors or zero-shot large models, SG-Bot exemplifies lightweight, real-time, and user-controllable characteristics, seamlessly blending the consideration of commonsense knowledge with automatic generation capabilities. SG-Bot employs a three-fold procedure--observation, imagination, and execution--to adeptly address the task. Initially, objects are discerned and extracted from a cluttered scene during the observation. These objects are first coarsely organized and depicted within a scene graph, guided by either commonsense or user-defined criteria. Then, this scene graph subsequently informs a generative model, which forms a fine-grained goal scene considering the shape information from the initial scene and object semantics. Finally, for execution, the initial and envisioned goal scenes are matched to formulate robotic action policies. Experimental results demonstrate that SG-Bot outperforms competitors by a large margin.