Check patentability & draft patents in minutes with Patsnap Eureka AI!

Image feature recognition method and device based on deep learning model and storage medium

A deep learning and image feature technology, applied in the field of artificial intelligence, can solve problems such as low efficiency and rely too much on human experience, and achieve the effect of improving accuracy and good performance

Pending Publication Date: 2020-12-11
SHENZHEN ZTE NETVIEW TECH +1
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to provide an image feature recognition method, device and storage medium based on a deep learning model, which solves the existing problem of relying on manual adjustment of hyperparameters of the pooling layer, which is inefficient and relies too much on human experience

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 feature recognition method and device based on deep learning model and storage medium
  • Image feature recognition method and device based on deep learning model and storage medium
  • Image feature recognition method and device based on deep learning model and storage medium

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach

[0050] In this embodiment, the number of parallel pooling layers in the pooling module is two or more, and the hyperparameters of each pooling layer are different, and the hyperparameters of each pooling layer must satisfy the formula (1). In a specific implementation, if the feature vector to be pooled is figure 2 In B, the pooling module includes three parallel pooling layers, and the hyperparameters of the three pooling layers are shown in Table 1.

[0051] Table 1

[0052] pooling kernel size step value zero padding 3x3 2 0 5x5 2 1 7x7 2 2

[0053] Input the feature vector B to be pooled into the three pooling layers shown in Table 1, and all three pooling layers can output a 3×3 sub-pooling feature vector.

[0054] In an embodiment, the pooling feature vector is obtained according to at least two of the sub-pooling feature vectors, including:

[0055] Adding at least two sub-pooling feature vectors element-wise to obtain a pooling ...

Embodiment 2

[0063] Please refer to image 3 , image 3 It is a structural block diagram of an image feature recognition device based on a deep learning model in an embodiment, and the image feature recognition device includes: an image acquisition module 101 , a feature extraction module 102 and a feature recognition module 103 .

[0064] The image acquisition module 101 is used to acquire image information to be identified. The image information in this embodiment may be a picture, and the picture has target features to be identified.

[0065] The feature extraction module 102 is used to input the image information to be identified into a pre-built deep learning model to obtain a feature vector; wherein, the pre-built deep learning model includes a pooling module, and the pooling module includes at least two pooling layers juxtaposed , the pooling module is used to input the feature vectors to be pooled received at its input into each pooling layer to perform pooling operations to obta...

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 feature recognition method and device based on a deep learning model, and a storage medium. The method comprises the steps that to-be-recognized image information is input into a pre-built deep learning model which comprises a pooling module, to-be-recognized image information is subjected to convolution, a to-be-pooled feature vector is input into a pooling module, the pooling module comprises at least two parallel pooling layers, each pooling layer performs pooling operation on the to-be-pooled feature vector to obtain at least two sub-pooling feature vectors, and the pooling feature vectors can be obtained according to the at least two sub-pooling feature vectors; according to the method, the pooling feature vectors are acquired and output for subsequentlearning to extract the corresponding feature vectors, the feature vectors to be pooled are subjected to pooling operation of different pooling layers, so that the feature vectors to be pooled can acquire richer information, no extra parameters are brought through the different pooling layers, the deep learning model has better performance, and the accuracy of target feature recognition in the image is improved.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to an image feature recognition method, device and storage medium based on a deep learning model. Background technique [0002] When identifying image features, it is necessary to extract the feature vectors in the acquired image information first. Generally, the acquired pictures are input into the deep learning model, and after operations such as convolution pooling, the deep learning model outputs a one-dimensional feature vector or a two-dimensional feature vector. Dimensional feature vector, the feature vector is the feature vector to be extracted, and the target feature is identified according to the feature vector. A deep learning model is generally composed of a convolutional layer, a pooling layer, a nonlinear layer, and a fully connected layer. The pooling layer in the deep learning model can reduce the size of the input image on the one hand, and o...

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): G06K9/46G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06V10/44G06N3/045G06F18/241
Inventor 李一力张浩邵新庆刘强徐明
Owner SHENZHEN ZTE NETVIEW TECH
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com