Improved Human-Object Interaction Based Lightweight Openpose

In order to improve the performance as well as the requirement of lightweight, the human-object interaction detection network of GGNet including the improved Lightweight Openpose is proposed. The Lightweight Openpose algorithm is introduced to substitute for the feature extraction module in the GGNe...

Celý popis

Uložené v:
Podrobná bibliografia
Vydané v:2022 International Conference on 3D Immersion, Interaction and Multi-sensory Experiences (ICDIIME) s. 104 - 107
Hlavní autori: Yang, Hanmin, Liu, Qingyuan, Zhang, Jiali
Médium: Konferenčný príspevok..
Jazyk:English
Vydavateľské údaje: IEEE 01.06.2022
Predmet:
Feature extraction
ggnet
human-object interaction
lightweight openpose
Task analysis
Three-dimensional displays
On-line prístup:Získať plný text
Tagy: Pridať tag
Žiadne tagy, Buďte prvý, kto otaguje tento záznam!
Popis
Shrnutí:In order to improve the performance as well as the requirement of lightweight, the human-object interaction detection network of GGNet including the improved Lightweight Openpose is proposed. The Lightweight Openpose algorithm is introduced to substitute for the feature extraction module in the GGNet network as the feature extraction network. The improved GGNet network has an accuracy of 96.6% and a recall of 82.6%, which is a significant improvement in speed and accuracy compared to using the original GGNet feature extraction network.
DOI:10.1109/ICDIIME56946.2022.00030