Convolutional network three-dimensional model reconstruction method based on multi-view cost volume
A technology of convolutional network and 3D model, applied in the field of 3D model reconstruction of convolutional network based on multi-view cost volume, can solve the problems of unsatisfactory realization effect and susceptibility to the influence of multiple modes, so as to improve rationality and guarantee irrelevant effect
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment
[0081] The objective tasks of the present invention are as follows figure 2 , image 3 , Figure 4 and Figure 5 as well as Image 6 shown, figure 2 is the input multi-view image, image 3 is the specific information of the camera parameters, Figure 4 is the voxel quantized representation for supervision, Figure 5 is the voxel visualization representation for supervision, Image 6 For the voxel results reconstructed by the network, the structure of the whole method is as follows Figure 7 shown. Each step of the present invention will be described below according to examples.
[0082] In step (1), a feature map of the input multi-view image data is extracted through a weight-sharing encoding network. Taking 3 perspectives as an example, it is divided into the following steps:
[0083] Step (1.1), the input dataset (taken from the 13 categories of the ShapeNet dataset) has a total of 20 rendering images for each sample (the format is .png, and the width, height a...
PUM
Abstract
Description
Claims
Application Information
- R&D Engineer
- R&D Manager
- IP Professional
- Industry Leading Data Capabilities
- Powerful AI technology
- Patent DNA Extraction
Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.
© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com