Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-view three-dimensional model retrieval method based on block convolutional neural network

A convolutional neural network and three-dimensional model technology, applied in the field of multi-view three-dimensional model retrieval based on block convolutional neural network, can solve the problems of large amount of calculation, unsatisfactory effect, and restrict the development of traditional methods, and achieve the effect of excellent performance.

Active Publication Date: 2020-10-02
SHANDONG ARTIFICIAL INTELLIGENCE INST +1
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, in practical applications, using traditional methods for rendering and feature extraction of 3D models is not only computationally intensive but also unsatisfactory.
These difficulties limit the development of related traditional methods

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-view three-dimensional model retrieval method based on block convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0025] Attached below figure 1 The present invention will be further described.

[0026] A multi-view 3D model retrieval method based on block convolutional neural network, comprising:

[0027] a) Rendering the 3D model to obtain N 2D views of the 3D model. The three-dimensional model may be a three-dimensional model generated by a computer, and the computer renders the three-dimensional model to obtain multiple two-dimensional views. The 3D model can also be a real-world object, and X cameras with different angles are set around the object to obtain two-dimensional views of the object at different angles. Preferably, a camera is set every 30 degrees in the circumferential direction around the object, and 12 two-dimensional views are obtained by setting 12 cameras.

[0028] b) The previous multi-view method directly inputs the obtained multiple views into the weight-shared convolutional neural network at the same time to obtain the view features of each view. In this metho...

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

No PUM Login to View More

Abstract

The invention discloses a multi-view three-dimensional model retrieval method based on a block convolutional neural network, and the method comprises: mining an internal relation between views in a view feature extraction process through employing a block convolution layer for a multi-view image; and according to the cosine similarity between each view feature and the maximum pooled view feature,allocating different weights to each view, and obtaining more differentiated model features by using the discrimination between the view features. When the loss function is generated, model features and view features are both considered, and the network can be better restrained for learning. The multi-view three-dimensional model retrieval method based on the block convolutional neural network achieves excellent performance in a related three-dimensional model retrieval data set.

Description

technical field [0001] The invention relates to the field of three-dimensional vision, in particular to a multi-view three-dimensional model retrieval method based on a block convolutional neural network. Background technique [0002] With the development of 3D representation technology and computer hardware performance, 3D vision has attracted more and more attention from researchers. Compared with traditional two-dimensional images, three-dimensional vision is a more realistic description of the real world, including the spatial structure information of three-dimensional objects and the characteristics of three-dimensional geometry and contour curves. 3D model retrieval is a research hotspot in the field of 3D vision. Related research methods can be divided into two stages (1) 3D model retrieval methods based on traditional methods, and (2) 3D model retrieval methods based on deep learning. [0003] The 3D model retrieval method based on the traditional method is to gene...

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
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06F16/53
CPCG06F16/53G06N3/045G06F18/213G06F18/24147G06F18/253G06F18/214
Inventor 高赞邵煜翔程志勇陈达舒明雷聂礼强
Owner SHANDONG ARTIFICIAL INTELLIGENCE INST
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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