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

A Support Vector Machine-Based Optimization Method for Key Parameters of Point Cloud Compression Encoder

A support vector machine and key parameter technology, applied in the field of image processing, to achieve the effect of improving quality and strong generalization ability

Active Publication Date: 2022-05-27
SHANDONG UNIV
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] Aiming at the optimization problem of encoding parameters in the existing PCL-PCC point cloud encoder, the present invention provides an encoder parameter optimization method based on support vector machine SVM, which greatly improves the compression performance of point cloud

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
  • A Support Vector Machine-Based Optimization Method for Key Parameters of Point Cloud Compression Encoder
  • A Support Vector Machine-Based Optimization Method for Key Parameters of Point Cloud Compression Encoder
  • A Support Vector Machine-Based Optimization Method for Key Parameters of Point Cloud Compression Encoder

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0110] A method for optimizing key parameters of point cloud compression encoder based on support vector machine, such as image 3 shown, including the following steps:

[0111] (1) The preprocessing process of point cloud data is as follows:

[0112] ① Homogenize the geometric information in the point cloud information; refer to: read the point cloud The geometric information (x, y, z) of all points in the point cloud, and the minimum value of the point cloud geometric information (x min , y min ,z min ) and the maximum value of the point cloud geometric information (x max , y max ,z max ).

[0113] ②For point cloud and the i-th point p i The geometric coordinate information in (x i , y i ,z i ) for uniform processing; i∈{1, 2,...N}, N is the number of points in the point cloud; as shown in formula (I):

[0114]

[0115] In formula (I), Represents the i-th point p after normalization i geometric coordinate information.

[0116] 3. Convert the color infor...

Embodiment 2

[0134] According to a support vector machine-based point cloud compression encoder key parameter optimization method described in Embodiment 1, the difference is:

[0135] In step ⑤, the point set obtained by projecting to the X_Y plane i-th projection point S xy，i Represents the i-th point projected onto the X_Y plane, and the formulas are shown in formula (IV), formula (V), formula (VI), formula (VII), and formula (VIII):

[0136]

[0137]

[0138]

[0139]

[0140]

[0141] In the formula, p xy_i Representing a point cloud Projected to the position on the two-dimensional X_Y plane The set of points at , where n is p xy_i The number of points in the set, when n=1, it means one-to-one projection, when n>1, it means many-to-one mapping; represents the set p xy_i The average value of the luminance signal at all points in , represents the set p xy_i The average value of the red color-difference signal at all points in ; represents the set p xy_i ...

Embodiment 3

[0157] According to a support vector machine-based point cloud compression encoder key parameter optimization method described in Embodiment 1, the difference is:

[0158] Step ⑥, extract the point set on the X_Y plane The distribution eigenvectors of As shown in formula (XIX), formula (XX), formula (XXI), formula (XXII), formula (XXIII), formula (XXIV):

[0159]

[0160]

[0161]

[0162]

[0163]

[0164]

[0165] In the formula, N xy express The number of midpoints.

[0166] Step ⑥, extract the point set on the Y_Z plane The distribution eigenvectors of As shown in formula (XXV), formula (XXVI), formula (XXVII), formula (XXVIII), formula (XXIX), formula (XXX):

[0167]

[0168]

[0169]

[0170]

[0171]

[0172]

[0173] In the formula, N yz express The number of midpoints.

[0174] Step ⑥, extract the point set on the X_Z plane The distribution eigenvectors of As shown in formula (XXXI), formula (XXXII), formula (XXXI...

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 present invention relates to a key parameter optimization method of a point cloud compression encoder based on a support vector machine. First, the geometric information and color information of the point cloud are preprocessed; then, the feature vector of the point cloud is extracted; for a given target code rate, use the full search method to find the optimal parameter pair that minimizes the distortion, extract the optimal label under a given target bit rate for all point clouds in the training set, and write the optimal label information and feature vector information into the training Set, use the support vector machine and training set information to train the model, use the model to test the feature vector information in the test set, predict the optimal test label on the continuous domain, and obtain the optimal parameter pair of the test set. This method utilizes the distribution characteristics of the point cloud, uses the support vector machine method to train the optimal encoding parameter pair of the test point cloud, and greatly reduces the time cost while ensuring the encoding performance of the encoder under the condition of a given encoding bit rate.

Description

technical field [0001] The invention relates to a three-dimensional point cloud compression coding optimization method based on a support vector machine, and belongs to the technical field of image processing. Background technique [0002] With the improvement of the capability of 3D data acquisition equipment, 3D point cloud has become an effective way to express objects or scenes. Three-dimensional point cloud (point cloud data) includes a collection of many points, namely to describe objects in space, where N represents the total number of points, p i Represents the i-th point, i∈{1,…N}, each point is usually composed of geometric information (ie, the spatial coordinates of the point x, y, z, the origin of the coordinate system is usually set as the center point of the acquisition device) and attribute components ( It usually includes three chromaticity attributes of R, G, and B at the point, as well as color information, as well as reflectivity, normal vector, etc.). ...

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): G06T9/00G06N20/10
CPCG06T9/00G06N20/10
Inventor 刘祺元辉王韦韦刘昊
Owner SHANDONG UNIV