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

Image migration method based on mean value standard deviation

A mean standard deviation, image technology, applied in digital image processing, artificial intelligence and neural network intersection fields, can solve the problems of poor feature capture ability and poor applicability

Pending Publication Date: 2021-12-24
GUILIN UNIVERSITY OF TECHNOLOGY
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The applicability of the artistic features of the model extraction based on local features is poor, it has certain limitations, and the overall feature capture ability is poor

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
  • Image migration method based on mean value standard deviation
  • Image migration method based on mean value standard deviation
  • Image migration method based on mean value standard deviation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0016] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0017] The flow chart of the present invention is as figure 1 As shown, the image style transfer method based on the mean standard deviation algorithm, the method specifically includes the following process:

[0018] Step 1: Style Extraction Network

[0019] Design a feature space, which can be built on the convolution kernels of each network layer to store feature information on different convolution kernels.

[0020] The features of the convolution kernels of different network layers are combined to obtain rich and stable features. In the field of neural style transfer, by using deep neural networks trained for object recognition, we can operate in feature spaces to explicitly represent the high-level content of images.

[0021] Among them, l represents the lth network layer, Indicates the feature value of the content picture p at positi...

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 image migration method based on a mean value standard deviation, and the method constructs a feature space to store feature information of different filters, achieves different normalized statistics in different network layers, better obtains multi-scale and stable features, does not need to train real data, and can flexibly carry out style transfer. According to CNN theoretical analysis, feature information of high-level network reconstruction is extracted on the basis, and an FC layer and a soft-max layer are removed to improve operation efficiency. Experimental results show that the performance of the mean value standard deviation algorithm is superior to that of a Gram algorithm in the style transfer process, the distortion effect of style transfer is small, and the time operation efficiency is improved by about 30 times.

Description

technical field [0001] The invention belongs to the intersection fields of digital image processing, artificial intelligence and neural network. The subject content is an image style transfer method based on the mean standard deviation algorithm, which has small distortion effect of style transfer and high transfer operation efficiency. Background technique [0002] Image style transfer is to extract the unique style features of the style image, and transfer this feature to the content image, so that the feature maps of the two can be combined. In the process of style transfer, the generated feature map should faithfully represent the artistic features of the original style image, and also render the texture features generated by the combination of the content image and the style image. The goal of style transfer is to make the intermediate image consistent with the content image in content and consistent with the style image in style after multiple parameter adjustments. ...

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): G06T3/00G06T5/50G06N3/04G06N3/08
CPCG06T5/50G06N3/084G06T2207/20221G06N3/045G06T3/04
Inventor 叶汉民李志波蒲立力
Owner GUILIN UNIVERSITY OF TECHNOLOGY