Progressive Point Cloud Upsampling via Differentiable Rendering

Zhang, Pingping; Wang, Xu; Ma, Lin; Wang, Shiqi; Jiang, Jianmin

doi:10.1109/tcsvt.2021.3100134

Cited by 24 publications

(3 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…By considering the whole and the details of the point clouds, the network can produce better results. Zhang et al [37] designed a feature-expansion module. It can learn local and global point features via a down-feature operator and an up-feature operator.…”

Section: Global Guided Samplingmentioning

confidence: 99%

Cyclic Global Guiding Network for Point Cloud Completion

et al. 2022

View full text Add to dashboard Cite

The application of 3D scenes has gradually expanded in recent years. A 3D point cloud is unreliable when it is acquired because of the performance of the sensor. Therefore, it causes difficulties in utilization. Point cloud completion can reconstruct and restore sparse and incomplete point clouds to a more realistic shape. We propose a cyclic global guiding network structure and apply it to point cloud completion tasks. While learning the local details of the whole cloud, our network structure can play a guiding role and will not ignore the overall characteristics of the whole cloud. Based on global guidance, we propose a variety of fitting planes and layered folding attention modules to strengthen the local effect. We use the relationship between the point and the plane to increase the compatibility between the network learning and the original sparse point cloud. We use the attention mechanism of the layer overlay to strengthen the local effect between the encode and decode. Therefore, point clouds are more accurate. Our experiments indicate the effectiveness of our method on the ShapeNet, KITTI, and MVP datasets and are superior to other networks.

show abstract

Section: Global Guided Samplingmentioning

confidence: 99%

Cyclic Global Guiding Network for Point Cloud Completion

et al. 2022

View full text Add to dashboard Cite

show abstract

“…Another way to approach the sparseness of points is to upsample the point clouds. Zhang et al [19] present a progressive method for point cloud upsampling via differentiable rendering, which addresses the non-uniform point distribution within the point cloud and is capable of learning local and global point features to cope with non-uniform point distribution and outlier removal. Yu et al [20] present PU-Net, a data-driven point cloud upsampling approach on point patches capable of learning multi-level point features and expanding a set of points using a multi-branch convolution unit implicitly in feature space.…”

Section: Point Cloud Renderingmentioning

confidence: 99%

Real-Time Interpolated Rendering of Terrain Point Cloud Data

Kordež¹,

Marolt²,

Bohak³

2022

Preprint

View full text Add to dashboard Cite

Most real-time terrain point cloud rendering techniques do not address the empty space between the points but rather try to minimize it by changing the way how the points are rendered by either rendering them bigger or with more appropriate shapes such as paraboloids. In this work, we propose an alternative approach to point cloud rendering, which addresses the empty space between the points and tries to fill it with appropriate values to achieve the best possible output. The proposed approach runs in real time and outperforms several existing point cloud rendering techniques in terms of speed and render quality.

show abstract

“…Zhang et al. [ 19 ] present a progressive method for point cloud upsampling via differentiable rendering, which addresses the non-uniform point distribution within the point cloud and is capable of learning local and global point features to cope with non-uniform point distribution and outlier removal. Yu et al.…”

Section: Related Workmentioning

confidence: 99%

Real-Time Interpolated Rendering of Terrain Point Cloud Data

Kordež

Marolt

Bohak

2022

Sensors

View full text Add to dashboard Cite

Most real-time terrain point cloud rendering techniques do not address the empty space between the points but rather try to minimize it by changing the way the points are rendered by either rendering them bigger or with more appropriate shapes such as paraboloids. In this work, we propose an alternative approach to point cloud rendering, which addresses the empty space between the points and tries to fill it with appropriate values to achieve the best possible output. The proposed approach runs in real time and outperforms several existing point cloud rendering techniques in terms of speed and render quality.

show abstract

Progressive Point Cloud Upsampling via Differentiable Rendering

Cited by 24 publications

References 48 publications

Cyclic Global Guiding Network for Point Cloud Completion

Cyclic Global Guiding Network for Point Cloud Completion

Real-Time Interpolated Rendering of Terrain Point Cloud Data

Real-Time Interpolated Rendering of Terrain Point Cloud Data

Contact Info

Product

Resources

About