Open Access

Object Localization Algorithm Based on Meta-Reinforcement Learning

Han Yan

Han Yan

Xi’an Technological University, School of Computer Science & EngineeringXi’an, China
Search for this author on
Sciendo|Google Scholar
Yan, Han
 and 
Hong Jiang

Hong Jiang

Xi’an Technological University, School of Computer Science & EngineeringXi’an, China
Search for this author on
Sciendo|Google Scholar
Jiang, Hong
  
Mar 16, 2024
Object Localization Algorithm Based on Meta-Reinforcement Learning's Cover Image
About this article
Previous Article
Next Article
Cite
Download Cover
Published Online: Mar 16, 2024
Page range: 55 - 65
DOI: https://doi.org/10.2478/ijanmc-2023-0077
Keywords
© 2023 Han Yan et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1.

Process of target localization in meta-reinforcement learning
Process of target localization in meta-reinforcement learning

Figure 2.

Object localization model
Object localization model

Figure 3.

Action diagram
Action diagram

Figure 4.

Structure of the location network
Structure of the location network

Figure 5.

Structure of the regression network
Structure of the regression network

Figure 6.

Feature network structure
Feature network structure

Figure 7.

Training process of meta-parameter pooling
Training process of meta-parameter pooling

Figure 8.

Parameter training process of meta-reinforcement learning
Parameter training process of meta-reinforcement learning

Figure 9.

Comparison of training loss functions
Comparison of training loss functions

Figure 10.

Results of ta
Results of ta

Figure 11.

Comparison of precision and recall
Comparison of precision and recall
Language:
English
Publication timeframe:
4 times per year
Journal Subjects:
Computer Sciences, Computer Sciences, other
Journal RSS Feed