Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Image style recognition method and device, computer equipment and storage medium

A recognition method and a technology of a recognition device, which are applied in the field of image recognition, can solve problems such as difficult to be widely used, low efficiency of mathematical or statistical models, loss, etc.

Pending Publication Date: 2020-08-11
SHENZHEN BINCENT TECH
View PDF3 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0002] In the prior art, image style recognition adopts manual feature extraction or common cnn classification model to extract image features for recognition. However, manually extracting image texture features as image style representation and building a mathematical or statistical model for style is too inefficient. Low, and it is difficult to be widely used, and the high-level extracted by the CNN classification model often includes a combination of abstract features such as color, texture, style, etc., and these features are highly abstract and will lose many details, resulting in low performance in image style recognition

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 style recognition method and device, computer equipment and storage medium
  • Image style recognition method and device, computer equipment and storage medium
  • Image style recognition method and device, computer equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0026] Embodiment 1 of the present invention provides an image style recognition method, such as figure 1 , image recognition methods include:

[0027] Step S11. Build a convolutional neural network, and input the picture to be recognized into the convolutional neural network to obtain m convolutional feature maps, where m is greater than 1.

[0028] In step S11, building a convolutional neural network refers to building a vgg16 model with 5 sets of convolutional layers, and inputting the image to be recognized into the vgg16 model to obtain m convolutional feature maps.

[0029] Among them, "input the picture to be recognized into the vgg16 model to obtain m convolutional feature maps" includes:

[0030] Input the picture to be recognized into the vgg16 model to obtain 512 14×14 convolution feature maps.

[0031] Such as figure 2 As shown, input a 224×224×3 picture, after two convolutions of 64 convolution kernels, use one pooling (pooling) to get a 112×112×64 picture; an...

Embodiment 2

[0045] Embodiment 2 of the present invention provides an image style recognition device, such as Figure 6 As shown, the image style recognition device includes:

[0046] The feature map acquisition module is used to build a convolutional neural network, and input the image to be recognized into the convolutional neural network to obtain m convolutional feature maps, where m is greater than 1;

[0047] The matrix acquisition module is used to compress each convolutional feature map into an n-dimensional vector to obtain m n-dimensional vectors, and perform inner product calculation on every two n-dimensional vectors to obtain an m×m Gram matrix;

[0048] The classification module is used for classifying the m×m Gram matrix and outputting the image style of the image to be recognized.

[0049] Further, the feature map acquisition module is specifically used to build a vgg16 model with 5 sets of convolution layers, and input the image to be recognized into the vgg16 model to ob...

Embodiment 3

[0053] Embodiment 3 of the present invention provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the elevator projection control method in the above-mentioned embodiment is implemented. In order to avoid repetition, it is not repeated here Let me repeat. Alternatively, when the computer program is executed by the processor, the functions of the modules / units in the elevator projection control device in the above-mentioned embodiments are realized, and details are not repeated here to avoid repetition.

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 provides an image style recognition method, an image style recognition device, computer equipment and a storage medium. The image style recognition method comprises the steps of: establishing a convolutional neural network, inputting a to-be-recognized image into the convolutional neural network to obtain m convolution feature maps, wherein m is greater than 1; compressing each convolution feature map into n-dimensional vectors to obtain m n-dimensional vectors, and performing inner product calculation on every two n-dimensional vectors to obtain an m*m Gram matrix; and inputtingthe m*m Gram matrix into a shallow network for classification, and outputting an image style of the to-be-recognized image. According to the image style recognition method and the image style recognition device, an intermediate feature map is extracted as a style feature, style features better than those of a full connection layer are obtained, the correlation features of the feature maps are calculated, the shallow network is used for learning the features, better style representation is obtained, the classification performance is improved, end-to-end learning is achieved in the process, andthe image style recognition method is simpler and more convenient than manual feature extraction.

Description

technical field [0001] The present invention relates to the technical field of image recognition, in particular to an image style recognition method, device, computer equipment and storage medium. Background technique [0002] In the prior art, image style recognition adopts manual feature extraction or common cnn classification model to extract image features for recognition. However, manually extracting image texture features as image style representation and building a mathematical or statistical model for style is too inefficient. Low, and it is difficult to be widely used, and the high-level extracted by the CNN classification model often includes a combination of abstract features such as color, texture, style, etc., and these features are highly abstract and will lose many details, resulting in low performance in image style recognition . Contents of the invention [0003] The object of the present invention is to provide an image style recognition method, device, ...

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): G06K9/62G06N3/04G06F17/16
CPCG06F17/16G06N3/045G06F18/241
Inventor 王国彬胡鹏侯兴兴
Owner SHENZHEN BINCENT TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com