View source for 18496063. METHOD AND APPARATUS FOR LEARNING DEPENDENCY OF FEATURE DATA simplified abstract (ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE)

=METHOD AND APPARATUS FOR LEARNING DEPENDENCY OF FEATURE DATA=


==Organization Name==

[[:Category:ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE|ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE]]


[[Category:ELECTRONICS AND TELECOMMUNICATIONS RESEARCH INSTITUTE]]

==Inventor(s)==

[[:Category:Jaehwan Kim of Daejeon (KR)|Jaehwan Kim of Daejeon (KR)]][[Category:Jaehwan Kim of Daejeon (KR)]]

[[:Category:Jung Jae Yu of Daejeon (KR)|Jung Jae Yu of Daejeon (KR)]][[Category:Jung Jae Yu of Daejeon (KR)]]

[[:Category:Wonyoung Yoo of Daejeon (KR)|Wonyoung Yoo of Daejeon (KR)]][[Category:Wonyoung Yoo of Daejeon (KR)]]

[[:Category:Juwon Lee of Daejeon (KR)|Juwon Lee of Daejeon (KR)]][[Category:Juwon Lee of Daejeon (KR)]]

==METHOD AND APPARATUS FOR LEARNING DEPENDENCY OF FEATURE DATA - A simplified explanation of the abstract==


This abstract first appeared for US patent application 18496063 titled 'METHOD AND APPARATUS FOR LEARNING DEPENDENCY OF FEATURE DATA

==Simplified Explanation==

The patent application describes a neural network device that learns the dependency of feature data for human body motion using graph information and self-attention mechanisms.

* The processor acquires graph information with data nodes for the human body.
* Feature data corresponding to multiple joints of the human body is extracted from the graph information.
* A self-attention output is obtained based on a self-attention mechanism.
* Result data for human body motion is generated using the self-attention output, which includes position information from positional encoding and structural information from geodesic encoding of the feature data.

== Potential Applications ==
This technology could be applied in various fields such as sports training, physical therapy, animation, and virtual reality development.

== Problems Solved ==
This innovation helps in accurately analyzing and predicting human body motion, which can be challenging due to the complex interactions between different joints and body parts.

== Benefits ==
- Improved understanding of human body movement
- Enhanced motion prediction and analysis
- Potential for personalized training programs

== Potential Commercial Applications ==
The technology could be utilized in sports performance analysis software, medical rehabilitation devices, animation software, and virtual reality applications.

== Possible Prior Art ==
One possible prior art could be existing motion capture systems that track human body movements for various applications.

=== Unanswered Questions ===

=== How does this technology compare to traditional motion capture systems in terms of accuracy and efficiency? ===
This article does not provide a direct comparison between this technology and traditional motion capture systems. Further research or testing may be needed to determine the performance differences.

=== What are the potential limitations or challenges in implementing this technology in real-world applications? ===
The article does not address any potential limitations or challenges that may arise when implementing this technology. Additional studies or practical applications could reveal obstacles that need to be overcome for successful deployment.


==Original Abstract Submitted==

A neural network device for learning dependency of feature data includes: a memory in which at least one program is stored; and a processor that performs a calculation by executing the at least one program, in which the processor is configured to acquire graph information including a data node for a human body; extract feature data corresponding to a plurality of joints constituting the human body from the graph information; acquire a self-attention output corresponding to the feature data based on a self-attention mechanism; and generate result data for a motion of the human body based on the self-attention output, and the self-attention output includes position information acquired based on positional encoding of the feature data and structural information acquired based on geodesic encoding of the feature data.



[[Category:G06V10/82]]
[[Category:G06T7/73]]
[[Category:G06V10/77]]