What question did this study set out to answer?

To develop a model that better recognizes human actions using skeletal data by incorporating intrinsic topology and temporal dynamics.

Intrinsic Topology and Multi-Scale Temporal Modelling for Skeleton-Based Human Action Recognition in Smart Surveillance Systems

View Full Paper

Ask AI

Bookmark

Authors

Innovative model integrates intrinsic topology and multi-scale dynamics for action recognition, enhancing surveillance efficacy.

To develop a model that better recognizes human actions using skeletal data by incorporating intrinsic topology and temporal dynamics.
Integrated intrinsic bone structure with multi-scale temporal dynamics for action recognition.
Utilized a topological space graph convolution module with a multi-head self-attention mechanism.
Developed a joint-bone interaction bridge for efficient skeletal data fusion and transmission.
Tested on NTU-RGB+D 60 and 120 datasets to evaluate performance.
Achieved 91.5% accuracy (CS) and 96.9% accuracy (CV) on NTU-RGB+D 60 dataset.
Attained 89.0% accuracy (C-Sub) and 90.8% accuracy (C-Set) on NTU-RGB+D 120 dataset.
Demonstrated superior feature extraction of spatiotemporal characteristics from skeletal data.

Khan et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

View Full Paper

Ask AI

Bookmark

Authors

Innovative model integrates intrinsic topology and multi-scale dynamics for action recognition, enhancing surveillance efficacy.

To develop a model that better recognizes human actions using skeletal data by incorporating intrinsic topology and temporal dynamics.
Integrated intrinsic bone structure with multi-scale temporal dynamics for action recognition.
Utilized a topological space graph convolution module with a multi-head self-attention mechanism.
Developed a joint-bone interaction bridge for efficient skeletal data fusion and transmission.
Tested on NTU-RGB+D 60 and 120 datasets to evaluate performance.
Achieved 91.5% accuracy (CS) and 96.9% accuracy (CV) on NTU-RGB+D 60 dataset.
Attained 89.0% accuracy (C-Sub) and 90.8% accuracy (C-Set) on NTU-RGB+D 120 dataset.
Demonstrated superior feature extraction of spatiotemporal characteristics from skeletal data.

Khan et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark