PATS: Proficiency-Aware Temporal Sampling for Multi-View Sports Skill Assessment

PATS (Proficiency-Aware Temporal Sampling) is a novel video sampling strategy designed specifically for automated sports skill assessment. Unlike traditional methods that randomly sample frames or use uniform intervals, PATS preserves complete fundamental movements within continuous temporal segments. The paper presenting PATS has been accepted at the 2025 4th IEEE Sport Technology and Research Workshop.

This tool showcases the PATS sampling strategy. Find out more at the project page: https://edowhite.github.io/PATS

Core Concept

The key insight is that athletic proficiency manifests through structured temporal patterns that require observing complete, uninterrupted movements. PATS addresses this by:

  • Extracting continuous temporal segments rather than isolated frames
  • Preserving natural movement flow essential for distinguishing expert from novice performance
  • Distributing multiple segments across the video timeline to maximize information coverage

Performance

When applied to SkillFormer on the EgoExo4D benchmark, PATS achieves:

  • Consistent improvements across all viewing configurations (+0.65% to +3.05%)
  • Substantial domain-specific gains: +26.22% in bouldering, +2.39% in music, +1.13% in basketball

Sampling Parameters
1 50
1 20
0.5 10