For athletes
Make form visible.
Compare movement over time, connect cues with video evidence, and understand what a coach or clinician is seeing.
Computer vision · biomechanics · coaching
See movement more clearly.
AI-assisted form analysis can turn ordinary video into estimated joint positions, timing, and movement patterns. The useful result is not a score—it is a better question for an athlete, coach, clinician, or researcher to examine.
AI estimates do not diagnose injury or replace a qualified coach or clinician. They can support consistent observation, comparison, and informed review.
What the work connects.
BYU’s Data-Driven Engineering in Athletics project brings engineering, AI, sports biomechanics, exercise science, data visualization, and community outreach into one workflow.
- Capture
- Authorized smartphone or field video is checked for camera position, frame rate, visibility, completeness, and suitability for analysis.
- Estimate
- Computer-vision and pose-estimation methods identify body landmarks and derive movement measurements for running, jumping, throwing, basketball, and related actions.
- Validate
- Researchers compare the estimates with biomechanical principles, reference measurements, and the original footage, while documenting uncertainty and failure cases.
- Translate
- Video overlays, reports, dashboards, and coaching summaries turn technical measurements into observations a qualified person can review and act on.
Performance and injury-risk reduction need context.
Movement data can help reveal asymmetry, timing changes, technique drift, or loading patterns that deserve attention. A responsible workflow treats those signals as prompts for human examination—not as automatic conclusions.
For coaches
Add measurements to observation.
Use repeatable video and data practices to support—not replace—sport knowledge, athlete communication, and professional judgment.
For researchers
Study reliability and transfer.
Test how camera conditions, model choices, athlete variation, and measurement uncertainty affect the validity of AI-assisted analysis.
Two connected entry points.
IRSRI presents the athlete-facing vision; BYU PRISM documents the research, student experience, methods, collaborators, and progress.
IRSRI
IRSRI describes a collaboration between BYU Engineering and Intermountain Health focused on expert biomechanics and clinical care, with smartphone video and AI-driven pose estimation intended to provide actionable feedback.
Open IRSRI ↗BYU PRISM
The Data-Driven Engineering in Athletics project details responsible AI practices, video quality assurance, computer vision, biomechanics, coach-facing communication, and interdisciplinary student research.
Open project page ↗Data-driven coaching course in development.
The course is being shaped around the whole chain from capture to communication, rather than a single app or model.
Video and data quality
Camera geometry, frame rate, visibility, consent, identifiers, secure handling, and the conditions under which analysis should stop.
Course moduleAI-assisted form analysis
Pose estimation, derived measurements, validation, uncertainty, and comparison with anatomy and biomechanical principles.
Course moduleCoach-facing interpretation
Turning measurements into careful observations, useful questions, video overlays, reports, and conversations that avoid overstating evidence.
Course moduleHealth and safety boundary
This material is educational. AI-based movement estimates may support a review process, but they are not medical advice and should not be used alone to diagnose, clear, or treat an athlete.