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

Image coloring method based on improved deep separable convolutional neural network

A convolutional neural network and image coloring technology, applied in the field of image processing technology and deep learning, can solve the problems of memory consumption and memory cost increase, restricting industrial applications, and low network computing efficiency, so as to reduce the amount of parameters and improve coloring Accuracy, the effect of reducing the number of parameter calculations

Pending Publication Date: 2021-03-12
NANJING UNIV OF SCI & TECH
View PDF2 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the currently proposed models often use more convolutional structures and deeper network layers. Although these model structures achieve better coloring performance, they often cause a sharp increase in memory consumption and memory costs, making network computing Low efficiency, not suitable for real-time processing, restricting the application in the industrial field

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 coloring method based on improved deep separable convolutional neural network
  • Image coloring method based on improved deep separable convolutional neural network
  • Image coloring method based on improved deep separable convolutional neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0024] Image coloring methods based on deep learning have achieved certain results in recent years, but most of them require a lot of computing resources and computing time, making it difficult to deploy on mobile or embedded devices. The present invention proposes an efficient method using a small number of parameters while substantially not affecting the coloring effect. The network structure of the present invention comprehensively considers global semantic features and local pixel features, and uses methods such as residual, depth-separable convolution, and channel weighting to reduce parameters and improve performance.

[0025] combine figure 1 , an image colorization method based on an improved depthwise separable convolutional neural network, including the following steps:

[0026] Step 1, construct an image dataset;

[0027] Step 2, constructing an improved depthwise separable convolutional coloring neural network; combined with figure 2 , the construction process ...

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 coloring method based on an improved depth separable convolutional neural network. The method comprises the following steps: constructing an image data set; constructing an improved deep separable convolutional coloring neural network; training an improved deep separable convolutional coloring neural network; and inputting the grayscale image to be colored into thetrained lightweight colored neural network to obtain an image colorization result. According to the network structure of the invention, global semantic features and local pixel features are comprehensively considered, and residual errors, depth separable convolution, channel weighting and other modes are used to reduce parameters and improve performance.

Description

technical field [0001] The invention belongs to the field of image processing technology and deep learning, in particular to an image coloring method based on an improved deep separable convolutional neural network. Background technique [0002] The human eye is far less sensitive to grayscale images than to color images. When the difference between adjacent pixels of the grayscale image is small, the human eye will not be able to capture the specific details in the image. Compared with grayscale images, color images can show richer environmental information and detailed textures of objects, which is more conducive to the extraction of image features by algorithms. Therefore, the efficient colorization of grayscale images has been paid more and more attention in the field of computer vision. [0003] Grayscale image coloring algorithms are mainly divided into three types, namely the traditional image coloring method based on coloring line expansion, the image coloring meth...

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): G06T11/40G06N3/04
CPCG06T11/40G06N3/045
Inventor 徐昱琨王清华李振华
Owner NANJING UNIV OF SCI & TECH
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