New System for Recognizing Actions of ADHD Children
Author Information
Author(s): Tang Wanyu, Shi Chao, Li Yuanyuan, Tang Zhonglan, Yang Gang, Zhang Jing, He Ling
Primary Institution: College of Biomedical Engineering, Sichuan University
Hypothesis
Can a novel keypoints-based system improve action recognition for children with ADHD during the TOVA assessment?
Conclusion
The Multi-cue Feature Fusion Network (MF-Net) achieved 90.6% accuracy in recognizing actions of ADHD children during the TOVA assessment.
Supporting Evidence
- The system achieved 90.6% top-1 accuracy and 97.6% top-2 accuracy across six action categories.
- The study involved 3801 video samples of ADHD children.
- The proposed MF-Net integrates features from both body and facial keypoints.
Takeaway
This study created a new system to help recognize how children with ADHD act during a test, making it easier for doctors to understand their behavior.
Methodology
The study used a Multi-cue Feature Fusion Network (MF-Net) to analyze video samples of ADHD children during the TOVA test, extracting features from body and facial keypoints.
Limitations
The study may have limitations related to the dataset size and the specific context of the TOVA assessment.
Participant Demographics
Children with ADHD, specifically during the TOVA assessment.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website