CVDec 18, 2023

Machine Vision-Enabled Sports Performance Analysis

arXiv:2312.11340v1h-index: 14
Originality Incremental advance
AI Analysis

It addresses the need for accessible, low-cost motion analysis tools for athletes and coaches, though it is incremental as it builds on existing markerless capture methods.

This study evaluated monocular 2D markerless motion capture using a smartphone for sports performance metrics, finding excellent agreement with ground truth for jump height and velocity but poor to moderate performance for other measures like flight time and range of motion.

$\textbf{Goal:}$ This study investigates the feasibility of monocular 2D markerless motion capture (MMC) using a single smartphone to measure jump height, velocity, flight time, contact time, and range of motion (ROM) during motor tasks. $\textbf{Methods:}$ Sixteen healthy adults performed three repetitions of selected tests while their body movements were recorded using force plates, optical motion capture (OMC), and a smartphone camera. MMC was then performed on the smartphone videos using OpenPose v1.7.0. $\textbf{Results:}$ MMC demonstrated excellent agreement with ground truth for jump height and velocity measurements. However, MMC's performance varied from poor to moderate for flight time, contact time, ROM, and angular velocity measurements. $\textbf{Conclusions:}$ These findings suggest that monocular 2D MMC may be a viable alternative to OMC or force plates for assessing sports performance during jumps and velocity-based tests. Additionally, MMC could provide valuable visual feedback for flight time, contact time, ROM, and angular velocity measurements.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes