Anisotropic convolution-based point cloud completion method and device
An anisotropic and convolutional technology, applied in the field of deep learning, can solve the problems of point cloud information loss, incomplete consideration of cross-point connectivity and context of adjacent points, etc.
Pending Publication Date: 2021-09-10
ZHEJIANG UNIV OF TECH
View PDF0 Cites 3 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0004] The purpose of this application is to provide a point cloud completion method and device based on anisotropic convolution, which overcomes the fact that the prior art does not fully consider the connectivity across points and the context of adjacent points when processing point clouds, thus The problem of point cloud information loss caused by
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0031] In order to make the purpose, technical solution and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described here are only used to explain the present application, not to limit the present application.
[0032] This application proposes a new network framework to predict missing point clouds from missing point clouds in two stages. In the first stage, a novel multi-resolution anisotropic convolutional encoder is adopted to better extract latent features of 3D objects from missing point clouds. These latent features contain not only local and global features, but also Low-level features and high-level features. In the second stage, a novel decoder is adopted to better infer missing point clouds from feature maps.
[0033] Such as figure 1 As shown, this application proposes a point cloud...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses an anisotropic convolution-based point cloud completion method and device. The method comprises the steps of: carrying out down-sampling of input original point cloud data by employing an iterative farthest point sampling algorithm, and obtaining two pieces of point cloud data with smaller resolution; inputting the two pieces of point cloud data into an anisotropic convolution encoder, and extracting the potential features of the point cloud data; connecting the three extracted potential features, and generating a final feature vector by using MLP; and inputting the final feature vector into a rough-to-fine combined pyramid decoder, and speculating a point cloud missing part from the final feature vector. According to the invention, more high-precision point clouds can be generated, and the distortion of the whole point clouds and the point clouds of missing areas is smaller.
Description
technical field [0001] The present application belongs to the technical field of deep learning, and in particular relates to a point cloud completion method and device based on anisotropic convolution. Background technique [0002] With the rapid development of UAV and satellite technology, the number of remote sensing images is increasing by tens of thousands. As a ground feature closely related to human life, buildings occupy a large proportion in remote sensing images. Therefore, building segmentation technology in high-resolution remote sensing images has always been the focus of research, which is of great significance for urban planning, land protection, and urban and rural renovation. The characteristics of the remote sensing image itself also make this segmentation extremely difficult. First of all, there are many occlusion and shadow problems in remote sensing images, and these unfavorable factors affect the judgment of building segmentation algorithms. Secondly,...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Applications(China)
IPC IPC(8): G06F17/15G06K9/46G06K9/62
CPCG06F17/153G06F18/214
Inventor 刘盛李丁达曹益峰黄文豪陈胜勇
Owner ZHEJIANG UNIV OF TECH