A Field Model for Repairing 3D Shapes

By Dinh C. Nguyen, Binh-Son Hua, Minh-Khoi Tran, Quang-Hieu Pham, Sai-Kit Yeung. 2016

122 citations recorded by the source corpus.

This paper proposes a field model for repairing 3D shapes constructed from multi-view RGB data. Specifically, we represent a 3D shape in a Markov random field (MRF) in which the geometric information is encoded by random binary variables and the appearance information is retrieved from a set of RGB images captured at multiple viewpoints. The local priors in the MRF model capture the local structures of object shapes and are learnt from 3D shape templates using a convolutional deep belief network. Repairing a 3D shape is formulated as the maximum a posteriori (MAP) estimation in the corresponding MRF. Variational mean field approximation technique is adopted for the MAP estimation. The proposed method was evaluated on both artificial data and real data obtained from reconstruction of practical scenes. Experimental results have shown the robustness and efficiency of the proposed method in repairing noisy and incomplete 3D shapes.

Subfields: Computational Mechanics, Computer Vision and Pattern Recognition, Geology

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Vision and Pattern Recognition

In Geology

Citations in the corpus

Learning Shape Priors for Single-View 3D Completion and Reconstruction (2018, positive): “[50] extend the method for mesh repairing.”
Shape Completion Using 3D-Encoder-Predictor CNNs and Shape Synthesis (2016, methodology): “Nguyen et al. [27] build on this work and apply it to repairing meshes; they use the input of 3D ShapeNets and compute a distance transform on which they apply a Markov Random Field.”
Semantic Scene Completion from a Single Depth Image (2016, background): “ited to address only part of this problem as shown in Figure 2: RGB-D segmentation approaches consider only visible surfaces without the full 3D shape [6, 26], while shape completion approaches consider only geometry without semantics [3] or a single object out of context [32, 34].”
VConv-DAE: Deep Volumetric Shape Learning Without Object Labels (2016, background): “[9] is a concurrent work that uses 3D to 2D(multi-view images) correspondence to repair 3D shapes.”
High-Resolution Shape Completion Using Deep Neural Networks for Global Structure and Local Geometry Inference (2017, background): “Nguyen et al. [18] combines CDBNs and Markov Random Fields (MRFs) to formulate shape completion as a Maximum a Posteriori (MAP) inference problem.”
GRNet: Gridding Residual Network for Dense Point Cloud Completion (2020, methodology): “Since the convolution can not be directly applied to point clouds due to their irregularity and unorderedness, most of the existing methods [2,3,4,5,6,7,8] voxelize the point cloud into binary voxels, where 3D convolutional neural networks can be applied.”
PCN: Point Completion Network (2018, methodology): “For example, surface reconstruction methods [3, 10, 32, 40, 48, 53, 61] generate smooth interpolations to fill holes in locally incomplete scans.”
SketchyGAN: Towards Diverse and Realistic Sketch to Image Synthesis (2018, methodology): “Similar distance-field representations are used in recent works on 3D shape recovery [51, 18].”
JSIS3D: Joint Semantic-Instance Segmentation of 3D Point Clouds With Multi-Task Pointwise Networks and Multi-Value Conditional Random Fields (2019, background): “, image structures, can be easily extended to 3D, which leads to deep learning with volumetric representation [47, 26, 28].”
ScanComplete: Large-Scale Scene Completion and Semantic Segmentation for 3D Scans (2017, background): “This network can generate and complete shapes, and also repair broken meshes [23].”
SceneNN: A Scene Meshes Dataset with aNNotations (2016, methodology): “It also allows the creation of object datasets [31].”
JSNet: Joint Instance and Semantic Segmentation of 3D Point Clouds (2019, background): “Literature research shows that 3D scene data have different representations, e.g., volumetric grids (Wu et al. 2015; Thanh Nguyen et al. 2016; Maturana and Scherer 2015) and 3D point clouds (Qi et al. 2017b; Li et al. 2018; Wang et al. 2018b; Yu et al. 2018).”
SnowflakeNet: Point Cloud Completion by Snowflake Point Deconvolution with Skip-Transformer (2021, background): “(1) Traditional point cloud completion methods [42, 1, 44, 26] usually assume a smooth surface of 3D shape, or utilize large-scaled complete shape dataset to infer the missing regions for incomplete shape.”