Unlock instant, AI-driven research and patent intelligence for your innovation.

3D Keypoint Detection Method Based on Deep Neural Network and Sparse Autoencoder

A sparse autoencoder and deep neural network technology, applied in the field of 3D key point detection based on deep neural network and sparse autoencoder, can solve the problem of lack of global information

Active Publication Date: 2019-07-19
UNIV OF ELECTRONICS SCI & TECH OF CHINA
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] However, most of the above algorithms only use local information to generate feature attributes, lacking global information such as Laplacian spectrum

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
  • 3D Keypoint Detection Method Based on Deep Neural Network and Sparse Autoencoder
  • 3D Keypoint Detection Method Based on Deep Neural Network and Sparse Autoencoder
  • 3D Keypoint Detection Method Based on Deep Neural Network and Sparse Autoencoder

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0063] The method of the present invention will be further described in detail below with reference to the accompanying drawings and specific examples. The goal of the example is to verify the effectiveness of the method of the present invention through the key point detection results of the three-dimensional mesh model.

[0064] In the implementation process, we used the literature (Dutagaci, H., Cheung, CP, Godil, A.: Evaluation of 3d interest point detection techniques via human-generated ground truth. The Visual Computer 28(9)(2012)901-917) The 3D grid model database is used as the training and testing data set.

[0065] The specific implementation of the training deep neural network stage:

[0066] Step 1. Select training set and test set from the 3D grid model database, and select positive and negative sample points from the training set:

[0067] The three-dimensional grid model database is divided into two parts: database A and database B. Database A contains 24 three-dimensio...

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 belongs to the technical field of three-dimensional computer vision, and in particular relates to a three-dimensional key point detection method based on a deep neural network and a sparse autoencoder. The method includes a stage of training a sparse autoencoder and a deep neural network and a stage of detecting three-dimensional key points by using the trained deep neural network as a regression model. The local and global information of the 3D mesh model in the multi-scale space is fully utilized to detect whether the point to be measured is a key point. Introducing a multi-layer sparse autoencoder can effectively discover the correlation between these local and global information and form a high-level feature representation of this information for regression on it. Finally, the key points in the 3D mesh model can be detected effectively, robustly and stably.

Description

Technical field [0001] The invention belongs to the technical field of three-dimensional computer vision, and specifically relates to a three-dimensional key point detection method based on a deep neural network and a sparse autoencoder. Background technique [0002] Three-dimensional key point detection is an important part of three-dimensional computer vision, which is widely used in various applications such as target registration and matching, three-dimensional shape retrieval, mesh segmentation and simplification. Researchers have proposed a variety of methods for detecting three-dimensional key points in the past few decades, most of which are based on geometry. Godila and Wagan extended the two-dimensional Scale Invariant Feature Transform (SIFT) algorithm and proposed a three-dimensional SIFT key point detection algorithm. Holte uses the Difference-of-Normals (DoN) operator to detect three-dimensional key points. Based on the principle of visual saliency of the 3D mesh ...

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 Patents(China)
IPC IPC(8): G06T17/30
CPCG06T17/30G06T2207/20081
Inventor 朱策林薪雨张倩王征韬刘翼鹏夏志强虢齐
Owner UNIV OF ELECTRONICS SCI & TECH OF CHINA