The myhairline.ai benchmark database already contains density curves for over 50,000 individual treatment courses, making it one of the largest real-world hair density datasets ever assembled. Every anonymized data point contributed by users helps refine AI accuracy for the entire community.
This guide explains what the benchmark database is, how your data is anonymized, what research it enables, and why collective contribution produces better outcomes for individual users.
What the Benchmark Database Contains
The benchmark database is a structured collection of anonymized hair density measurements, treatment records, and classification data. It does not contain photos, names, email addresses, or any information that could identify an individual user.
Each contributed record includes:
| Data Field | Example | Purpose |
|---|---|---|
| Density readings (follicular units per cm2) | 145 FU/cm2 | Trains density estimation models |
| Norwood classification | Stage 3V | Refines classification accuracy |
| Treatment type | Finasteride 1mg daily | Builds treatment response curves |
| Treatment duration | 8 months | Maps response timelines |
| Density change over time | +12% at 6 months | Calculates expected outcomes |
| Age range | 30-35 | Adjusts predictions by age |
| Ethnicity category | Asian | Calibrates ethnicity-specific norms |
| Donor zone density | 170 FU/cm2 | Informs extraction planning |
No individual record can be traced back to a specific person. The anonymization process removes all metadata, geolocation, device identifiers, and timestamps that could enable re-identification.
Why Benchmark Data Matters for AI Accuracy
AI density estimation models are only as good as their training data. A model trained on 500 Caucasian males performs poorly on Asian hair types. A model trained on only Norwood 3 to 5 stages misclassifies early Stage 2 and late Stage 6 patterns.
The benchmark database solves this problem through volume and diversity. With over 50,000 treatment courses spanning every Norwood stage, multiple ethnicities, and dozens of treatment combinations, the AI produces accurate readings regardless of the user's demographic profile.
Ethnicity-Specific Calibration
Hair density varies significantly by ethnicity. Without benchmark data from diverse populations, AI models default to Caucasian norms and produce inaccurate readings for other groups.
| Ethnicity | Average Follicular Units per cm2 | Benchmark Contribution |
|---|---|---|
| Caucasian | 200 | Well-represented |
| Asian | 170 | Growing rapidly |
| African | 150 | Actively recruiting |
| Hispanic | 170 | Growing steadily |
| Middle Eastern | 180 | Growing steadily |
Every contribution from underrepresented demographics has an outsized impact on AI accuracy for that group.
Treatment Response Curves
One of the most valuable outputs of the benchmark database is treatment response curves. These curves show the expected density change over time for specific treatments, broken down by Norwood stage, age, and other variables.
For example, the database shows that finasteride (1mg daily) halts further loss in 80 to 90% of users and produces regrowth in approximately 65%. But the benchmark data reveals more granular patterns: Norwood 2 and 3 users see faster response than Norwood 5 and 6 users. Users under 30 respond more strongly than users over 50.
These nuances only emerge from large datasets. No single clinical trial captures this level of detail across real-world conditions.
How Anonymization Works
Data anonymization follows a multi-step process designed to make re-identification statistically impossible.
Step 1: Identifier Removal. All account-linked identifiers (user ID, email hash, session tokens) are permanently deleted from the contributed record.
Step 2: Generalization. Precise values are generalized into ranges. A user aged 33 becomes "30 to 35." A density reading taken on February 15, 2026 becomes "Q1 2026." A location in Chicago becomes "North America."
Step 3: Aggregation. Individual records are never exposed directly. All queries against the benchmark database return aggregated results across a minimum of 50 records. No researcher or API user can access a single individual's data.
Step 4: Differential Privacy. Statistical noise is added to query results. This mathematical guarantee ensures that adding or removing any single person's data does not meaningfully change the output of any query.
What Research the Database Enables
Norwood Progression Modeling
By tracking density changes over time across thousands of users, the benchmark database reveals average progression rates by Norwood stage. This data helps predict how quickly a Stage 3 user might progress to Stage 4, given their current rate of density decline.
This predictive capability helps users and surgeons plan interventions at the optimal time, not too early (wasting resources) and not too late (when donor supply may be insufficient for the required graft count).
Treatment Combination Effectiveness
Real-world data captures treatment combinations that clinical trials rarely study. The benchmark database tracks users on finasteride alone, minoxidil alone, both together, PRP in combination with medication, and dozens of other combinations.
PRP therapy alone produces a 30 to 40% density increase at $500 to $2,000 per session. But how does PRP perform when added to an existing finasteride regimen? The benchmark database answers questions like this with real-world density data.
Post-Transplant Graft Survival Benchmarks
FUE procedures achieve 90 to 95% graft survival rates in clinical literature. The benchmark database validates this with real-world tracking data and breaks it down by surgeon, technique, and patient demographics.
Users tracking their post-transplant recovery contribute data points that help future patients set realistic expectations. If the benchmark shows that FUE recovery takes 7 to 10 days on average but users over 50 take 10 to 14 days, that insight benefits every future patient.
How Your Contribution Helps You Personally
Contributing to the benchmark database is not just altruistic. Your own tracking results improve when the AI improves.
More accurate density readings. Every contribution refines the density estimation model. Your next reading benefits from the combined data of every other contributor.
Better treatment predictions. When the benchmark database shows that users with your demographic profile and Norwood stage respond to finasteride within 4 months, you can calibrate your expectations against thousands of similar cases.
Refined Norwood classification. Classification accuracy improves with more training examples. Borderline cases between Stage 2 and 3, or Stage 4 and 5, become easier for the AI to resolve correctly with more benchmark data.
Contribution Scale and Impact
The value of the benchmark database scales non-linearly. The first 1,000 contributors established baseline patterns. The next 10,000 refined those patterns into reliable predictions. At 50,000+ treatment courses, the database produces statistically significant insights for increasingly narrow demographic subgroups.
| Benchmark Size | Capability Level |
|---|---|
| 1,000 records | Basic Norwood classification |
| 5,000 records | Treatment response curves by stage |
| 10,000 records | Ethnicity-specific calibration |
| 25,000 records | Age and ethnicity combined modeling |
| 50,000+ records | Granular subgroup predictions |
Each new contribution, even a single user tracking for 3 months, adds another data point that strengthens predictions for everyone who shares a similar profile.
Opt-In Process and Controls
Data contribution is entirely voluntary. Here is how the opt-in process works:
- Default state: Off. No data is shared unless you explicitly enable contribution.
- Enabling contribution: Toggle "Contribute to Benchmark Database" in your account settings.
- What gets shared: Only anonymized density readings, classifications, and treatment logs.
- What never gets shared: Photos, personal identifiers, exact dates, or precise locations.
- Disabling contribution: Toggle off at any time. Previously contributed data remains anonymized in the aggregate pool but no new data is sent.
- Deletion requests: You can request deletion of your anonymized contribution, though individual records may already be aggregated beyond separation.
Data Security and Compliance
The benchmark database follows privacy standards including GDPR, CCPA, and BIPA requirements for biometric data handling. All data is encrypted at rest and in transit. Access to the raw benchmark database is restricted to internal AI training pipelines and authorized research partners.
No benchmark data is sold to third parties. Research partnerships require ethics board approval and produce only aggregated, published findings that benefit the broader hair loss community.
Medical disclaimer: This article is for informational purposes only and does not constitute medical advice. Treatment response data from the benchmark database represents population averages and may not predict individual outcomes. Consult a qualified medical professional for personalized treatment decisions.
Ready to contribute to the science of hair loss while tracking your own progress? Start your free analysis at myhairline.ai/analyze and join the largest real-world hair density benchmark in the world.