A Video-Based Classification System for Assessing Locomotor Skills in Children

Area of Science:

• Pediatric kinesiology
• Motor development assessment
• Biomechanical analysis

Background:

• The Test of Gross Motor Development 2 (TGMD-2) is the standard for assessing fundamental movement skills (FMS).
• TGMD-2's large-scale application is limited by its inefficiency and need for expert training.
• There is a need for more efficient and accessible methods to assess children's motor skills.

Purpose of the Study:

• To evaluate the accuracy of a novel video-based classification system (VCS) using a marker-less sensor for assessing children's locomotor skills.
• To compare the VCS's performance against the established TGMD-2 assessment.
• To determine if VCS can serve as a viable alternative for FMS evaluation.

Main Methods:

• 203 typically-developing children (3-8 years) performed six locomotor skills per TGMD-2 guidelines.
• Kinect v2 sensor captured movements; videos were rated by a trained rater and the VCS.
• Computational-kinematic algorithms were developed for instant performance rating by the VCS.

Main Results:

• VCS showed moderate-to-very good agreement with the rater (66.1%-87.5% per skill, 72.4% descriptive).
• No significant differences were found between TGMD-2 and VCS standard scores, with a significant positive correlation.
• Tukey mean difference plot indicated no bias (mean difference -0.16 ± 1.8), and moderate kappa agreement (k=0.54) for descriptive ratings.

Conclusions:

• The video-based classification system (VCS) demonstrates potential as an alternative to the TGMD-2 for assessing children's locomotor skills.
• VCS offers a more efficient and accessible method, potentially eliminating the need for an experienced rater.
• This technology could facilitate broader and more frequent assessments of fundamental movement skills in children.