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

A Scalable Category-Based Image Recognition Approach Based on Plastic Convolutional Neural Networks

A convolutional neural network and neural network technology, applied in neural learning methods, biological neural network models, neural architectures, etc.

Active Publication Date: 2022-04-19
NAT UNIV OF DEFENSE TECH
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The technical problem to be solved by the present invention is: for the neural network method to perform picture recognition, it is necessary to obtain a large number of sample pictures of this category for training, and the algorithm can only identify a few fixed categories that participate in the training. The present invention provides a An image recognition method for scalable categories based on plastic convolutional neural network. This method adopts the combination of plastic network and convolutional neural network structure and the combination of loop judgment. It can recognize multiple Unfixed categories for image 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
  • A Scalable Category-Based Image Recognition Approach Based on Plastic Convolutional Neural Networks
  • A Scalable Category-Based Image Recognition Approach Based on Plastic Convolutional Neural Networks

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0023] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

[0024] figure 1 It is the structure diagram of the convolutional plastic neural network constructed in the first step of the present invention. The network consists of an 11-layer structure. The first and last layers are the input layer (receives sequentially input pictures) and the output layer (the output length is 5 encoding results), and the 2nd to 9th layers are composed of convolutional pooling layers alternately. The tenth layer is the classification layer constructed by the plastic network layer. The relevant parameters involved in each layer such as convolution kernel size, step size, etc. are in Figure 1 It has been marked and can be adjusted according to actual needs.

[0025] figure 2 It is a flow chart of the specific implementation of the recognition calculation in the present invention. Taking 5-way-1-shot as an example, t...

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 belongs to the field of image recognition, and discloses an image recognition method based on a plastic convolutional neural network with scalable categories. The present invention constructs a convolutional plastic neural network with the ability of meta-learning, and manages and establishes the corresponding class library to be identified according to the application requirements, combines the plastic neural network, uses the library to be identified as data support, and uses the loop judgment method to recognize the input image and return the recognition result. The present invention solves the scalability problem of the identifiable category of the traditional target recognition method, and its advantage is that when there is an unknown category to be identified, only one or several image samples of the unknown category need to be stored in the library of the category to be identified In , there is no need to retrain the neural network or adjust the recognition algorithm, and the demand for samples is smaller than that of traditional methods.

Description

technical field [0001] The invention belongs to the field of image recognition, and relates to a calculation method for target image recognition, in particular to a target recognition method with meta-learning ability combining a convolutional neural network and a plastic neural network, and using the method to realize scalable category image recognition. Background technique [0002] Image recognition technology and method is the product of the information industry in today's era, and it is also a very popular research direction in the field of computer vision and digital image processing research. Image recognition technology has a very important role in military, medical and civilian fields in the world today. Widely used in robot navigation systems, unmanned driving technology, intelligent video monitoring, industrial product inspection and production process monitoring, aerospace and many other fields, image recognition technology is indispensable. Therefore, image reco...

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 Patents(China)
IPC IPC(8): G06V10/74G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/22
Inventor 王戟杨文婧杨绍武黄达徐利洋黄万荣胡亚清刘向阳沙建松颜豪杰梁卓
Owner NAT UNIV OF DEFENSE TECH