Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Automatic three-dimensional model component category tagging method

A 3D model and component category technology, applied in computer parts, character and pattern recognition, image data processing, etc., can solve the problems of quality dependence of segmentation results, no use of boundary information, etc.

Active Publication Date: 2016-11-23
NANJING UNIV
View PDF3 Cites 16 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method only considers generating a patch classifier based on the patch label information in the training data, and does not use the boundary information given in the label data, but obtains a rough boundary by over-segmenting the target shape Position, and then optimized, the quality of the final segmentation results greatly depends on the results of over-segmentation

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
  • Automatic three-dimensional model component category tagging method
  • Automatic three-dimensional model component category tagging method
  • Automatic three-dimensional model component category tagging method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0070] A fast three-dimensional model part category automatic labeling method disclosed by the present invention specifically includes the following steps ( figure 1 is the flowchart of the whole method):

[0071] Step 1, rapid training of the 3D model annotation training set: the training set contains similar models and the parts in the model have been given standard annotations, where the annotation information includes the category annotation of each mesh patch in the 3D model belonging to the constituent parts and each Mesh edges are subject to category labeling of boundary edges. Such as Figure 2a As shown, it is the standard label of the patch. Figure 2b is the standard label for mesh edges. Each aircraft model mesh patch is marked by different label classes indicated by different colors, and the labels in the training set are standard labels. Through rapid training on the 3D model labeling training set, a fast patch labeling model and a boundary labeling model tha...

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 an automatic three-dimensional model component category tagging method. The method comprises the following steps that in the training process, quick training is conducted according to a three-dimensional model tagging training set, and a quick tagging model used for classifying and tagging surface patches and mesh edges of an unknown three-dimensional model is obtained through training; in the tagging process, the quick tagging model, obtained through training, of the surface patches and mesh edges is utilized for classifying surface patches and mesh edges of a target model, classification probability distribution of the surface patches and mesh edges is obtained, a graph model is constructed, smoothing and optimizing of segmentation boundaries are conducted through multi-tag graph cut optimization, and therefore rapid automatic target three-dimensional model component category tagging is achieved.

Description

technical field [0001] The invention relates to a processing method for shape analysis, which belongs to the technical field of computer graphics, in particular to an automatic labeling method for three-dimensional model component categories. Background technique [0002] Understanding the component composition of 3D models is the basis for shape understanding and high-level geometry manipulation. In recent years, researchers have gradually introduced mature machine learning techniques into 3D shape analysis research, and data-driven joint segmentation methods for geometric models are receiving more and more attention. Using the core concept of data-driven "information transfer", a correlation can be established between the input pre-specified or calculated sample shape and the target shape to be analyzed and processed, and the information of interest can be transferred from the sample shape to Passed to the target shape. Therefore, data-driven shape segmentation can produ...

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
IPC IPC(8): G06T19/20G06T7/00G06K9/62
CPCG06T19/20G06T2219/004G06T2207/20081G06T2207/20072G06T2207/20084G06F18/24
Inventor 孙正兴李红岩宋沫飞武蕴杰
Owner NANJING UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Eureka Blog
Learn More
PatSnap group products