kinemotion

Validation Status

Document Version: 1.0 Last Updated: 2025-01-13 Status: Pre-validation

Executive Summary

Kinemotion is a pre-validation tool that has not been validated against gold standard measurement systems (force plates, 3D motion capture). While the implementation is technically sophisticated and biomechanically sound, it lacks empirical validation required for research or clinical use.

Bottom Line: Use kinemotion for training monitoring and exploratory analysis. Do not use for research publications, clinical decisions, or any context requiring validated accuracy.

Current Validation Status

What Has Been Done ✅

✅ Algorithmic implementation - Biomechanically sound algorithms based on validated theory
✅ Code quality - 74.27% test coverage (261 tests), type-checked, linted
✅ Theoretical foundation - Based on validated principles:
- Flight-time method (Bosco et al., 1983)
- Force-velocity profiling theory (Samozino et al., 2014)
- Dempster body segment parameters (1955)
✅ Open source - Transparent algorithms, reproducible results

What Has NOT Been Done ❌

❌ Force plate comparison - No validation against gold standard
❌ Reliability studies - No test-retest, inter-device, or inter-rater data
❌ Accuracy metrics - No ICC, bias, limits of agreement calculated
❌ Correction factors - No systematic bias corrections (cf. Bishop et al., 2022 required correction factors)
❌ Peer review - No published validation study
❌ Clinical validation - No testing in clinical populations
❌ Multi-site validation - No testing across different laboratories/conditions

Comparison to Validated Tools

Aspect	MyJump (Validated)	Kinemotion (Unvalidated)
Force plate validation	Yes (ICC=0.997)	No
Peer-reviewed study	Published 2015	None
Sampling rate	120Hz (iPhone 5s)	30-60fps (video dependent)
Method	Manual frame selection	Automated landmark tracking
Accuracy vs force plate	Bias: 1.1±0.5cm	Unknown
Cost	~$12 app	Free (open source)
Expertise required	Low (frame selection)	None (fully automated)
Use in research	Yes (widely cited)	No (not validated)

Key Difference: MyJump has been validated against force plates and published in peer-reviewed journals. Kinemotion has not.

Current Validation Efforts (AthletePose3D)

We are currently in Phase 1 of a multi-stage validation study using the AthletePose3D (AP3D) dataset, which contains high-speed athletic movements captured by research-grade multi-camera systems.

🧪 Methodology Highlights

To transition from 2D pixel-based analysis to research-grade 3D kinematics, we are implementing:

Coordinate Normalization: Using Procrustes Analysis (Kabsch Algorithm) to align MediaPipe detections with physical world coordinates (mm).
3D Joint Angles: Reconstructing anatomical angles (Hip, Knee, Ankle) in 3D space to correct for camera perspective and foreshortening.

For detailed technical analysis of why these steps matter for monocular video and how they are implemented, see the AthletePose3D Methodology Guide.

What Kinemotion CAN Claim

Technical Implementation ✅

“Implements biomechanically sound algorithms”
“Uses validated theoretical frameworks (Samozino, Bosco)”
“Provides consistent, repeatable measurements” (deterministic)
“Advanced signal processing (RANSAC, bilateral filtering)”
“Open source with comprehensive test coverage”

Practical Use ✅

“Suitable for training monitoring (relative changes)”
“Useful for educational purposes”
“Can track progress within individual athletes”
“Provides exploratory analysis capabilities”

What Kinemotion CANNOT Claim

Validation Claims ❌

~~“Validated against force plates”~~
~~“Research-grade accuracy”~~
~~“Clinically validated”~~
~~“Equivalent to MyJump/validated apps”~~
~~“Suitable for research publications”~~

Accuracy Claims ❌

~~“Accuracy within X cm of force plates”~~
~~“ICC > 0.90 reliability”~~
~~“Gold standard measurement”~~
~~“Laboratory-grade precision”~~

Known Limitations

1. No Gold Standard Validation

Issue: Never compared against force plates or optical motion capture Impact: Unknown accuracy, cannot make quantitative accuracy claims Mitigation: Validation study planned (see roadmap)

2. MediaPipe Pose Estimation Constraints

Issue: MediaPipe validated for clinical joint angles (Armstrong et al., 2025) but NOT for jump performance metrics Limitations documented in research:

Affected by lighting conditions
Reduced accuracy with loose clothing
Jitter from self-occlusion
Not optimized for high-speed ballistic movements

Impact: Tracking quality varies with recording conditions Mitigation: Quality indicators and confidence scores (planned)

3. Lower Sampling Rate

Issue: Typical video (30-60fps) vs validated apps (120-240Hz) Impact: Lower temporal resolution may reduce accuracy for rapid movements Comparison:

MyJump validated: 120Hz
MyJumpLab validated: 240Hz
Kinemotion typical: 30-60fps (2-4× lower)

Mitigation: Sub-frame interpolation, recommend 60fps minimum

4. Indirect Measurement Method

Issue: Landmarks → CoM estimation → velocity → phase detection Impact: Error propagation through processing pipeline Comparison:

MyJump: Direct feet tracking → flight time
Kinemotion: Landmarks → CoM → kinematic calculation (more steps = more error)

Mitigation: Robust filtering, outlier detection, physics checks

5. No Systematic Bias Correction

Issue: Bishop et al. (2022) showed correction factors needed for time-to-take-off when comparing manual vs automated detection Impact: May have uncorrected systematic bias Example from Bishop 2022:

Raw TTTO had bias vs force plate
Correction equation: y = 0.8947x + 0.1507 eliminated bias

Mitigation: Correction factors will be derived after validation study

Use Cases & Recommendations

✅ Appropriate Use Cases

1. Training Monitoring

Track jump performance trends over time
Compare athlete’s current vs baseline performance
Monitor fatigue, training response, recovery
Why appropriate: Relative changes don’t require absolute accuracy

2. Educational Purposes

Learn biomechanics of jumping
Understand video analysis methods
Teach pose estimation technology
Why appropriate: Educational value independent of validation

3. Exploratory Research

Pilot studies before formal testing
Proof-of-concept investigations
Methods development
Why appropriate: Preliminary work, not final conclusions

4. Self-Monitoring Athletes

Personal training feedback
Technique analysis
Progress tracking
Why appropriate: Self-directed, low-stakes context

❌ Inappropriate Use Cases

1. Research Publications

Cannot cite as validated measurement tool
Reviewers will reject unvalidated instruments
Alternative: Use validated equipment or conduct validation first

2. Clinical Decision-Making

Injury assessment
Return-to-play clearance
Treatment effectiveness
Why inappropriate: Patient safety requires validated tools
Alternative: Use clinical-grade force plates or validated apps

3. Talent Identification

Draft combines
Team selection
Athletic scholarships
Why inappropriate: High-stakes decisions require validated accuracy
Alternative: Use force plates or validated commercial systems

4. Legal/Insurance Context

Disability assessment
Injury compensation
Litigation evidence
Why inappropriate: Legal scrutiny requires validated measurements
Alternative: Only use validated, certified equipment

5. Absolute Performance Comparisons

Comparing athletes against normative data
Setting performance standards
Ranking athletes
Why inappropriate: Unknown accuracy affects absolute values
Alternative: Within-athlete comparisons only, or use validated equipment

Validation Roadmap

Planned Validation Activities (3 Months)

Phase 1: Internal Consistency (Month 1)

Test-retest reliability (determinism test)
Parameter sensitivity analysis
Known height validation (dropped objects)
Confidence scoring implementation

Phase 2: Comparative Validation (Month 2)

Manual frame selection comparison (vs MyJump method)
Inter-device reliability study
Frame rate sensitivity analysis
Validation utilities module

Phase 3: Documentation & Publication (Month 3)

Physics-based sanity checks
Environmental condition testing
Technical validation report
Preprint publication (OSF/SportRxiv)

Expected Outcomes

Test-retest ICC calculated
Correlation with manual method (target: r>0.95)
Known limitations documented
Optimal recording conditions established

What This Will NOT Provide:

Force plate validation (requires lab access)
Peer-reviewed publication (requires longer timeline)
Clinical validation (requires patient populations)

Future Validation (When Lab Access Available)

Gold Standard Validation:

Force plate comparison study (n≥20 subjects)
Calculate ICC, bias, limits of agreement
Derive correction factors if needed
Publish in peer-reviewed journal

Target Metrics:

ICC > 0.90 (excellent reliability)
Bias < 2cm (per Balsalobre-Fernández et al., 2015)
95% of differences within ±5cm

For detailed roadmap, see docs/development/validation-roadmap.md.

Recommended Recording Setup

To Maximize Accuracy (Based on MediaPipe Literature)

Camera Setup:

Frame rate: 60fps minimum (higher is better)
Resolution: 1920×1080 or higher
Distance: 2-3 meters from athlete
Angle: Lateral view (sagittal plane), perpendicular to movement
Height: Tripod-mounted at mid-torso height
Stability: Fixed position (no handheld)

Environment:

Lighting: Bright, even lighting (avoid shadows)
Background: Plain, uncluttered background
Surface: Indoor gym preferred (consistent conditions)

Athlete:

Clothing: Fitted clothing (compression gear ideal)
Visibility: Ensure full body visible throughout jump
Framing: Capture from head to feet with margin

Avoid:

Dim lighting or backlighting
Loose, baggy clothing
Cluttered backgrounds
Handheld/moving camera
Partial occlusion of body
Very low frame rates (<30fps)

Frequently Asked Questions

Q: Can I use kinemotion for my research study?

A: Not as a validated measurement instrument. You can:

Use it for pilot/exploratory work
Compare it against validated equipment
Conduct your own validation study
Cite it as experimental software

Do not:

Claim validated accuracy
Use as primary outcome measure
Compare against normative data

Q: How does kinemotion compare to MyJump?

MyJump: Validated (ICC=0.997 vs force plate), published, widely used in research
Kinemotion: Unvalidated, experimental, not suitable for research

If you need validated measurements, use MyJump or similar validated apps.

Q: Why should I use kinemotion instead of MyJump?

A: You probably shouldn’t if you need validated accuracy. Use kinemotion if:

You want free, open-source software
You need fully automated analysis (no manual frame selection)
You’re doing exploratory work
You want to understand/modify the algorithms
You’re tracking relative changes (not absolute accuracy)

Q: Can I track my own training progress with kinemotion?

A: Yes! For personal use tracking relative changes over time, kinemotion is appropriate. Just understand:

Absolute jump height values may have unknown error
Relative changes (trends) more reliable than absolute values
Use consistent recording setup for comparability
Consider validation if making training decisions

Q: When will kinemotion be validated?

Short-term (3 months): Internal validation, manual comparison, technical report
Long-term (when lab access available): Force plate validation, peer-review publication
No timeline: Depends on resources and collaborations

Q: Can I cite kinemotion in my paper?

A: You can cite the software repository, but:

Cannot claim: “Validated measurement tool”
Can say: “Experimental video analysis software”
Must note: “Not validated against gold standard equipment”
Recommend: Validate yourself or use alongside validated instruments

Proper citation:

Kinemotion (version X.X.X) [Software]. (2025). Retrieved from https://github.com/KinemotionInc/kinemotion

Note: Kinemotion is experimental software not validated against force plates.
Measurements should be interpreted with caution.

Contact & Contributions

Report Issues

If you discover accuracy issues or unexpected behavior:

Open an issue: https://github.com/KinemotionInc/kinemotion/issues
Include video specifications, detection results, expected vs actual

Contribute to Validation

If you have access to validation equipment (force plates, motion capture):

Contact maintainers about collaboration
Share validation data (with appropriate permissions)
Contribute to validation roadmap

Stay Updated

Watch the repository for validation study updates
Check docs/validation/ for new validation reports
See CHANGELOG.md for validation-related improvements

References

Validation Literature

Balsalobre-Fernández et al. (2015) - MyJump validation against force plates (ICC=0.997)
Bishop et al. (2022) - MyJumpLab validation with correction factors
Armstrong et al. (2025) - MediaPipe Pose validation for clinical kinematics
Bland & Altman (1986) - Methods for assessing agreement between measurements

Biomechanics Theory

Bosco et al. (1983) - Flight-time method for jump height calculation
Samozino et al. (2014) - Force-velocity imbalance and jump performance
Dempster (1955) - Body segment parameters for biomechanical modeling

Methodological Standards

McGraw & Wong (1996) - Intraclass correlation coefficient calculation
Hopkins (2000) - Measures of reliability in sports medicine

Document Changelog

Version 1.0 (2025-01-13):

Initial validation status documentation
Comprehensive limitations disclosure
Use case recommendations
Validation roadmap overview

This site is open source. Improve this page.