SplAttN: Bridging 2D and 3D with Gaussian Soft Splatting and Attention for Point Cloud Completion

• URL: http://arxiv.org/abs/2605.01466v1
• Date: Sat, 02 May 2026 14:34:21 GMT
• Title: SplAttN: Bridging 2D and 3D with Gaussian Soft Splatting and Attention for Point Cloud Completion
• Abstract summary: We propose SplAttN, which replaces hard projection with Differentiable Gaussian Splatting to produce a dense, continuous image-plane representation.<n>Experiments show that SplAttN achieves state-of-the-art performance on PCN and ShapeNet-55/34.
• Score: 17.124815270085048
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Although multi-modal learning has advanced point cloud completion, the theoretical mechanisms remain unclear. Recent works attribute success to the connection between modalities, yet we identify that standard hard projection severs this connection: projecting a sparse point cloud onto the image plane yields an extremely sparse support, which hinders visual prior propagation, a failure mode we term Cross-Modal Entropy Collapse. To address this practical limitation, we propose SplAttN, which replaces hard projection with Differentiable Gaussian Splatting to produce a dense, continuous image-plane representation. By reformulating projection as continuous density estimation, SplAttN avoids collapsed sparse support, facilitates gradient flow, and improves cross-modal connection learnability. Extensive experiments show that SplAttN achieves state-of-the-art performance on PCN and ShapeNet-55/34. Crucially, we utilize the real-world KITTI benchmark as a stress test for multi-modal reliance. Counter-factual evaluation reveals that while baselines degenerate into unimodal template retrievers insensitive to visual removal, SplAttN maintains a robust dependency on visual cues, validating that our method establishes an effective cross-modal connection. Code is available at https://github.com/zay002/SplAttN.

Related papers

• Cross-Coordinate Correspondence Pruning for Image-to-Point Cloud Registration [70.44225088078274]
  We propose a novel Cross-Coordinate Correspondences Pruning (CCP) strategy to acquire sufficient inliers while ensuring a low outlier ratio.<n>Our method achieves a significant performance improvement, surpassing existing state-of-the-art methods by at least 8.6% in Registration Recall across various benchmarks.
  arXiv  Detail & Related papers  (2026-07-19T11:45:14Z)
• PUFM++: Point Cloud Upsampling via Enhanced Flow Matching [15.738247394527024]
  PUFM++ is an enhanced flow-matching framework for reconstructing point clouds from sparse, noisy, and partial observations.<n>We introduce a two-stage flow-matching strategy that first learns a direct, straight-path flow from sparse inputs to dense targets, and then refines it using noise-perturbed samples to approximate the terminal marginal distribution better.<n>Experiments on synthetic benchmarks and real-world scans show that PUFM++ sets a new state of the art in point cloud upsampling.
  arXiv  Detail & Related papers  (2025-12-24T06:30:42Z)
• Unsupervised Online 3D Instance Segmentation with Synthetic Sequences and Dynamic Loss [52.28880405119483]
  Unsupervised online 3D instance segmentation is a fundamental yet challenging task.<n>Existing methods, such as UNIT, have made progress in this direction but remain constrained by limited training diversity.<n>We propose a new framework that enriches the training distribution through synthetic point cloud sequence generation.
  arXiv  Detail & Related papers  (2025-09-27T08:53:27Z)
• Let Images Give You More:Point Cloud Cross-Modal Training for Shape Analysis [43.13887916301742]
  This paper introduces a simple but effective point cloud cross-modality training (PointCMT) strategy to boost point cloud analysis. To effectively acquire auxiliary knowledge from view images, we develop a teacher-student framework and formulate the cross modal learning as a knowledge distillation problem. We verify significant gains on various datasets using appealing backbones, i.e., equipped with PointCMT, PointNet++ and PointMLP.
  arXiv  Detail & Related papers  (2022-10-09T09:35:22Z)
• Self-Supervised Arbitrary-Scale Point Clouds Upsampling via Implicit Neural Representation [79.60988242843437]
  We propose a novel approach that achieves self-supervised and magnification-flexible point clouds upsampling simultaneously. Experimental results demonstrate that our self-supervised learning based scheme achieves competitive or even better performance than supervised learning based state-of-the-art methods.
  arXiv  Detail & Related papers  (2022-04-18T07:18:25Z)
• IDEA-Net: Dynamic 3D Point Cloud Interpolation via Deep Embedding Alignment [58.8330387551499]
  We formulate the problem as estimation of point-wise trajectories (i.e., smooth curves) We propose IDEA-Net, an end-to-end deep learning framework, which disentangles the problem under the assistance of the explicitly learned temporal consistency. We demonstrate the effectiveness of our method on various point cloud sequences and observe large improvement over state-of-the-art methods both quantitatively and visually.
  arXiv  Detail & Related papers  (2022-03-22T10:14:08Z)
• Continuity-Discrimination Convolutional Neural Network for Visual Object Tracking [150.51667609413312]
  This paper proposes a novel model, named Continuity-Discrimination Convolutional Neural Network (CD-CNN) for visual object tracking. To address this problem, CD-CNN models temporal appearance continuity based on the idea of temporal slowness. In order to alleviate inaccurate target localization and drifting, we propose a novel notion, object-centroid.
  arXiv  Detail & Related papers  (2021-04-18T06:35:03Z)
• Exploring Data Augmentation for Multi-Modality 3D Object Detection [82.9988604088494]
  It is counter-intuitive that multi-modality methods based on point cloud and images perform only marginally better or sometimes worse than approaches that solely use point cloud. We propose a pipeline, named transformation flow, to bridge the gap between single and multi-modality data augmentation with transformation reversing and replaying. Our method also wins the best PKL award in the 3rd nuScenes detection challenge.
  arXiv  Detail & Related papers  (2020-12-23T15:23:16Z)
• Pseudo-LiDAR Point Cloud Interpolation Based on 3D Motion Representation and Spatial Supervision [68.35777836993212]
  We propose a Pseudo-LiDAR point cloud network to generate temporally and spatially high-quality point cloud sequences. By exploiting the scene flow between point clouds, the proposed network is able to learn a more accurate representation of the 3D spatial motion relationship.
  arXiv  Detail & Related papers  (2020-06-20T03:11:04Z)
• StickyPillars: Robust and Efficient Feature Matching on Point Clouds using Graph Neural Networks [16.940377259203284]
  StickyPillars is a fast, accurate and extremely robust deep middle-end 3D feature matching method on point clouds. We present state-of-art art accuracy results on the registration problem demonstrated on the KITTI dataset. We integrate our matching system into a LiDAR odometry pipeline yielding most accurate results on the KITTI dataset.
  arXiv  Detail & Related papers  (2020-02-10T17:53:41Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.