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

A small target recognition method, device and medium based on super-resolution reconstruction

A super-resolution reconstruction and super-resolution technology, applied in character and pattern recognition, instruments, biological models, etc., can solve problems such as poor performance, and achieve the effect of sufficient feature extraction, easy recognition, and improved recognition performance.

Active Publication Date: 2022-04-12
HUAZHONG UNIV OF SCI & TECH
View PDF9 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] Aiming at the defects and improvement needs of the prior art, the present invention provides a small target recognition method, device and medium based on super-resolution reconstruction to solve the poor performance of the existing target recognizer on the small target recognition problem question

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 small target recognition method, device and medium based on super-resolution reconstruction
  • A small target recognition method, device and medium based on super-resolution reconstruction
  • A small target recognition method, device and medium based on super-resolution reconstruction

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0029] A small target recognition method based on super-resolution reconstruction, such as figure 1 shown, including:

[0030] S1. Build a recognition model including a generator and a discriminator;

[0031] S2. Taking the low-resolution image as the input of the generator, taking the super-resolution image as the output of the generator, taking the real high-resolution image or the super-resolution image as the input of the discriminator, and The input of the discriminator is the probability of the real high-resolution image and the recognition result as the output of the discriminator; the recognition model is trained in combination with the generator loss function and the discriminator loss function; wherein, the generator loss function L G Expressed as: L G =L MSE +0.006×L VGG +0.001×L ADV +β×L CLS , where L MSE , L VGG , L ADV , L CLS are the pixel mean square error loss, VGG feature matching loss, confrontation loss, and target recognition loss, and β is the ...

Embodiment 2

[0072] A small target recognition device based on super-resolution reconstruction, including:

[0073] The model construction module is used to construct a recognition model including a generator and a discriminator; wherein, the low-resolution image is used as the input of the generator, the super-resolution image is used as the output of the generator, and the real high-resolution image is used as the output of the generator. The image or the super-resolution image is used as the input of the discriminator, and the probability that the input of the discriminator is a real high-resolution image and the recognition result are used as the output of the discriminator;

[0074] The model training module is used to initialize the network parameters of the generator and the discriminator, and train the recognition model in combination with the generator loss function and the discriminator loss function; wherein, the generator loss function L G Expressed as: L G = L MSE +0.006×L ...

Embodiment 3

[0079] The embodiment of the present application also provides a computer-readable storage medium. One or more non-transitory computer-readable storage media containing computer-executable instructions that, when executed by one or more processors, cause the processors to perform small super-resolution-based reconstruction Steps of the object recognition method.

[0080] A computer program product comprising instructions, when run on a computer, causes the computer to perform a super-resolution reconstruction-based small object recognition method.

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 a small target recognition method, device and medium based on super-resolution reconstruction. The method includes: constructing a recognition model including a generator and a discriminator; The image is used as the output of the generator, the real high-resolution image or super-resolution image is used as the input of the discriminator, and the probability of the input of the discriminator as the real high-resolution image and the recognition result are used as the output of the discriminator, combined to generate The loss function of the discriminator and the loss function of the discriminator are used to train the recognition model; based on the particle swarm optimization algorithm, the F1 score of the trained model on the verification sample set is selected as the fitness function value, and the β value corresponding to the largest F1 score is used as the optimal weight Coefficient, thus determine the optimal recognition model, and based on the optimal recognition model for small target recognition. In this way, the recognition performance of the model for small targets can be effectively improved.

Description

technical field [0001] The present invention relates to the technical field of target recognition, and more specifically, to a small target recognition method, device and medium based on super-resolution reconstruction. Background technique [0002] In real life, as the application of electronic equipment becomes more and more common in social production and people's life, massive image data is generated all the time, and there are many small objects in the image data collected by various cameras. For example, some small ground targets captured by drones, some small-sized lesions in medical images, and small targets such as pedestrians and vehicles captured by surveillance cameras. Therefore, accurate recognition of small targets is very important for analyzing and processing these image data, and small target recognition technology is also of great value in medical image analysis, security systems, video surveillance and tracking, automatic driving and other fields, and bas...

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): G06V20/10G06V10/82G06T3/40G06N3/00G06N3/04
CPCG06T3/4053G06N3/006G06V20/10G06N3/045
Inventor 胡静陈智勇张旭阳沈宜帆熊涛张美琦张宏志
Owner HUAZHONG UNIV OF SCI & 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