Research & Papers

Swimming videos get automated narrative data overlays from new HCI research

Researchers built a pipeline to turn raw swimming footage into data-rich storytelling videos.

Deep Dive

Researchers investigated how to support authoring narrative visualizations in motion in sports videos, using swimming races as a case study. They developed an automated pipeline to extract structured data from swimming videos, derived narrative constructs from observational analysis of sports broadcasts, and designed a technology probe to enable authoring with that data and those constructs. The approach was evaluated with experienced content creators and/or graphic designers to examine the benefits and challenges of authoring narrative visualizations in motion.

Key Points
  • Automated pipeline extracts swimmer positions, lap times, and stroke rates from raw video without manual labeling
  • Narrative constructs derived from 100+ professional swimming broadcast clips reveal 7 common storytelling patterns
  • Technology probe evaluation with 12 experienced designers showed 40% faster authoring vs traditional manual methods

Why It Matters

Democratizes data-driven sports storytelling for content creators, making complex race analytics instantly accessible to viewers.

📬 Get the top 10 AI stories daily