Summary of multi-stream techniques utilized in human action recognition.

<p>Summary of multi-stream techniques utilized in human action recognition.</p>

Saved in:

Bibliographic Details
Main Author:	Mohammed H. Al-Hakimi (22404258) (author)
Other Authors:	Ibrar Ahmed (3299847) (author), Muhammad Haseeb (689140) (author), Taha H. Rassem Senior Member IEEE (22404261) (author), Fahmi H. Quradaa (17896527) (author), Rashad S. Almoqbily (17896530) (author)
Published:	2025
Subjects:	Neuroscience Biotechnology Cancer Biological Sciences not elsewhere classified Information Systems not elsewhere classified temporal dependencies critical propose two architectures based action recognition 60 benchmark demonstrate order temporal derivatives subtle motion dynamics offering structured representations level spatial normalization agcn attains 95 agcn achieves 96 xlink "> skeleton high order features motion features via order spatial representations fail motion streams human motion enhanced spatial stream agcn branch agcn weighted fusion surpassing state structure encoding structural relationships integrates joint independent branches grained actions fused adaptively findings confirm final layer early layer distinguishing fine discriminative patterns comprehensive experiments art methods anatomical distances 5 %,
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items