Citation

Chung, Jen Li and Ong, Lee Yeng and Leow, Meng Chew (2025) A Comparative Analysis of Skeleton-based Human Action Recognition on DOOH Advertising. In: 7th IEEE International Conference on Artificial Intelligence in Engineering and Technology, IICAIET 2025, 6 August 2025 - 28 August 2025, Kota Kinabalu.

Text IEEE Xplore Full-Text PDF_1.pdf - Published Version Restricted to Repository staff only Download (1MB)

Abstract

Digital Out-of-Home (DOOH) advertising is increasingly deployed in public places, offering dynamic and customizable content to attract diverse audiences. It is important to understand the audience response to help advertisers optimize their advertising strategies. However, the questionnaires and facial analysis used in existing studies are time-consuming, subjective, and prone to errors from occlusion or tracking failure. Skeleton-based human action recognition (HAR) provides a promising alternative by modelling human actions through posture and movement, thereby enabling analysis of audience response based on physical actions and dwell time. Despite its potential, skeleton-based HAR has not yet been applied in DOOH advertising. This study presents a comparative analysis of three skeleton-based HAR models, including single-stream convolutional neural network (1SCNN), three-stream CNN (3S-CNN), and spatial temporal graph convolutional network (ST-GCN). The analysis employed a subset of actions from the NTU RGB+D dataset, specifically selected for relevance to DOOH advertising. This subset comprises six actions commonly observed in public advertising scenarios. The experimental results show the STGCN achieved the highest recognition accuracy in crosssubject and cross-view evaluation protocols, whereas the 3SCNN demonstrated the fastest computational performance.

Item Type:	Conference or Workshop Item (Paper)
Uncontrolled Keywords:	human action recognition, convolutional neural network
Subjects:	Q Science > Q Science (General) > Q300-390 Cybernetics
Divisions:	Faculty of Information Science and Technology (FIST)
Depositing User:	Ms Rosnani Abd Wahab
Date Deposited:	17 Mar 2026 01:47
Last Modified:	17 Mar 2026 01:47
URII:	http://shdl.mmu.edu.my/id/eprint/15456

Downloads

Edit (login required)