A method for realizing image matrix convolution, computing equipment and storage medium
A technology of image matrix and computing equipment, which is applied in the field of data processing, can solve the problems of computing speed limitation, low efficiency, cumbersome implementation of image matrix convolution, etc., and achieve the effect of simple implementation, maximum computing efficiency, and improved convolution computing speed
Active Publication Date: 2021-10-15
成都统信软件技术有限公司
View PDF10 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, the above-mentioned image matrix convolution process is only based on the functions of the software, but only relies on the compiler to realize the image matrix convolution, and the calculation speed is limited. To improve the calculation efficiency, additional equipment needs to be connected, so that the existing The implementation of image matrix convolution is cumbersome and inefficient
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0062] An exemplary embodiment of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it is understood that the present disclosure can be implemented in various forms and should not be restricted herein. Instead, it is provided to provide more thoroughly understood the present disclosure, and can communicate the scope of the disclosure to those skilled in the art.
[0063] Image matrix volume is a tool that is very effective and quick to get image processing. In the image processing, the image is convolved in the form of a matrix, i.e., the image matrix is convolved, and the process of the image matrix convolution can be regarded as a "slip window" process, specifically, using a smaller matrix. As the convolutionary nucleus slides through the entirely convolved matrix, each sliding place requires a matrix internal capacity to make a matri...
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
Login to View More Abstract
The invention discloses a method for realizing convolution of an image matrix, a computing device and a storage medium. The method includes: obtaining an image matrix to be convoluted and a convolution kernel, and the matrix of the image to be convoluted is to convert the image to be convoluted according to the size of the convolution kernel. The matrix is expanded into a row matrix to obtain the first expanded image matrix, and the first expanded image matrix is converted into a column matrix to obtain the second expanded image matrix, and the convolution kernel is expanded to use the number of columns of the second expanded image matrix as the number of columns, The third expansion matrix is obtained by taking the size of the convolution kernel as the number of rows, the data size of each row in the second expansion image matrix, and the data size of each line in the third expansion matrix are the size of the vector register, and the second expansion image matrix Perform convolution operation with the third expansion matrix to obtain the feature matrix of the image matrix. The invention utilizes the vector register of the CPU in the calculation device to implement multiple floating-point data operations at the same time, and significantly improves the convolution calculation speed of the image matrix.
Description
Technical field [0001] Technical Field The present invention relates to data processing, and particularly relates to a method for implementing image convolution matrix, and computing device storage media. Background technique [0002] In the information age, the data processing (eg image data) requirements of real-time, efficiency is higher and higher, as a very basic image matrix convolution data processing means is in a critical position, thus improving the image matrix volume computational efficiency is the product of the main issues of current concern. [0003] Prior art, improve the computational efficiency multi-image matrix by convolution algorithms, e.g. im2col algorithm. im2col algorithm can be understood as a two-dimensional image matrix in advance according to the size of the convolution kernel, left to right, top to bottom in order to expand one of the rows, and the new line continuously distributed memory, rather to expand to the original matrix row matrix. When conv...
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
Login to View More Patent Type & AuthorityPatents(China)
IPC IPC(8): G06T1/60G06T1/20G06F17/16
CPCG06T1/60G06T1/20G06F17/16Y02D10/00
Inventor王正阳张勇刘明航
Owner成都统信软件技术有限公司