GenSDF: Two-Stage Learning of Generalizable Signed Distance Functions

• URL: http://arxiv.org/abs/2206.02780v1
• Date: Mon, 6 Jun 2022 17:58:29 GMT
• Title: GenSDF: Two-Stage Learning of Generalizable Signed Distance Functions
• Abstract summary: We investigate the generalization capabilities of neural signed distance functions (SDFs) for learning 3D object representations for unseen and unlabeled point clouds. Existing methods can fit SDFs to a handful of object classes and boast fine detail or fast inference speeds, but do not generalize well to unseen shapes. We introduce a two-stage semi-supervised meta-learning approach that transfers shape priors from labeled to unlabeled data to reconstruct unseen object categories.
• Score: 37.07344649854001
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: We investigate the generalization capabilities of neural signed distance functions (SDFs) for learning 3D object representations for unseen and unlabeled point clouds. Existing methods can fit SDFs to a handful of object classes and boast fine detail or fast inference speeds, but do not generalize well to unseen shapes. We introduce a two-stage semi-supervised meta-learning approach that transfers shape priors from labeled to unlabeled data to reconstruct unseen object categories. The first stage uses an episodic training scheme to simulate training on unlabeled data and meta-learns initial shape priors. The second stage then introduces unlabeled data with disjoint classes in a semi-supervised scheme to diversify these priors and achieve generalization. We assess our method on both synthetic data and real collected point clouds. Experimental results and analysis validate that our approach outperforms existing neural SDF methods and is capable of robust zero-shot inference on 100+ unseen classes. Code can be found at https://github.com/princeton-computational-imaging/gensdf.

Related papers

• ShapeGen: Robotic Data Generation for Category-Level Manipulation [57.10250198611112]
  ShapeGen is a method to generate 3D manipulation data in a simulator-free and 3D manner.<n>In the first stage, we train spatial warpings between shapes mapping points to points that correspond functionally, and aggregate 3D models along with the warpings into a plug-and-play Shape Library.<n>In the second stage, we design a pipeline that, leveraging established Libraries, requires only minimal human annotation to generate physically plausible and functionally correct demonstrations.
  arXiv  Detail & Related papers  (2026-04-16T22:55:39Z)
• Inferring Neural Signed Distance Functions by Overfitting on Single Noisy Point Clouds through Finetuning Data-Driven based Priors [53.6277160912059]
  We propose a method to promote pros of data-driven based and overfitting-based methods for better generalization, faster inference, and higher accuracy in learning neural SDFs. We introduce a novel statistical reasoning algorithm in local regions which is able to finetune data-driven based priors without signed distance supervision, clean point cloud, or point normals.
  arXiv  Detail & Related papers  (2024-10-25T16:48:44Z)
• A Lightweight UDF Learning Framework for 3D Reconstruction Based on Local Shape Functions [42.840655419509346]
  This paper presents a novel neural framework, LoSF-UDF, for reconstructing surfaces from 3D point clouds by leveraging local shape functions.<n>Our method exhibits enhanced resilience to noise and outliers in point clouds compared to existing methods.
  arXiv  Detail & Related papers  (2024-07-01T14:39:03Z)
• Class-Imbalanced Semi-Supervised Learning for Large-Scale Point Cloud Semantic Segmentation via Decoupling Optimization [64.36097398869774]
  Semi-supervised learning (SSL) has been an active research topic for large-scale 3D scene understanding. The existing SSL-based methods suffer from severe training bias due to class imbalance and long-tail distributions of the point cloud data. We introduce a new decoupling optimization framework, which disentangles feature representation learning and classifier in an alternative optimization manner to shift the bias decision boundary effectively.
  arXiv  Detail & Related papers  (2024-01-13T04:16:40Z)
• DiffComplete: Diffusion-based Generative 3D Shape Completion [114.43353365917015]
  We introduce a new diffusion-based approach for shape completion on 3D range scans. We strike a balance between realism, multi-modality, and high fidelity. DiffComplete sets a new SOTA performance on two large-scale 3D shape completion benchmarks.
  arXiv  Detail & Related papers  (2023-06-28T16:07:36Z)
• Visual Recognition with Deep Nearest Centroids [57.35144702563746]
  We devise deep nearest centroids (DNC), a conceptually elegant yet surprisingly effective network for large-scale visual recognition. Compared with parametric counterparts, DNC performs better on image classification (CIFAR-10, ImageNet) and greatly boots pixel recognition (ADE20K, Cityscapes)
  arXiv  Detail & Related papers  (2022-09-15T15:47:31Z)
• PatchRD: Detail-Preserving Shape Completion by Learning Patch Retrieval and Deformation [59.70430570779819]
  We introduce a data-driven shape completion approach that focuses on completing geometric details of missing regions of 3D shapes. Our key insight is to copy and deform patches from the partial input to complete missing regions. We leverage repeating patterns by retrieving patches from the partial input, and learn global structural priors by using a neural network to guide the retrieval and deformation steps.
  arXiv  Detail & Related papers  (2022-07-24T18:59:09Z)
• Implicit Shape Completion via Adversarial Shape Priors [46.48590354256945]
  We present a novel neural implicit shape method for partial point cloud completion. We combine a conditional Deep-SDF architecture with learned, adversarial shape priors. We train a PointNet++ discriminator that impels the generator to produce plausible, globally consistent reconstructions.
  arXiv  Detail & Related papers  (2022-04-21T12:49:59Z)
• Unsupervised Point Cloud Pre-Training via Occlusion Completion [18.42664414305454]
  We describe a simple pre-training approach for point clouds. It works in three steps: Mask all points occluded in a camera view; 2. Learn an encoder-decoder model to reconstruct the occluded points; 3. Use the encoder weights as initialisation for downstream point cloud tasks.
  arXiv  Detail & Related papers  (2020-10-02T16:43:14Z)

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.