KAI-VORAL-DSD-001 — AI Dataset Datasheet
Template ID: TPL-07-05 | Stage: 07-Test-and-Validate | Version: 0.3
Status: Project record created from the controlled template.
Purpose: Document a dataset used to train, tune or evaluate a model: provenance, rights, composition, labels, splits and limits.
Required inputs: Data inventory; consent/licence records; annotation protocol; split scripts.
Document control
| Field |
Required entry |
| Document ID |
KAI-VORAL-DSD-001 |
| Template used |
TPL-07-05 AI Dataset Datasheet v0.3 |
| Product / project / software version or commit SHA |
VORAL (Visionary Oral) / PRJ-001 / a6b4429e3165 (branch app_dev of AI-Dental-Project/Visionary_Oral) |
| Record status and version |
Draft v0.1 |
| Author / technical reviewer / quality reviewer |
Claude Code (drafted with Claude Code for KrinosAI) / Tech Lead (name to be assigned) / Clinical advisor (name to be assigned) |
| Created / updated / approved dates |
2026-10-28 / 2026-10-28 / Pending |
| Applicable SOP and revision |
SOP-LC-07 v0.3 |
| Related issue, commit, requirement and evidence links |
mobile/ml/model-manifest.json; mobile/ml/README.md; KAI-VORAL-MDC-001 |
| Approver and decision record |
Pending |
Applicability and evidence baseline
Product / system boundary: Web app and API (dentalapp/), doctor/admin mobile app (mobile/), on-device models (mobile/assets/models/), PDF Lambda (dentalapp/pdf-lambda/) at a6b4429e3165, and the external services they call.
Related approved requirements, risks and change records: None approved. Related drafts are listed in document control.
Retrospective record? Yes — follows the Retrospective Documentation Protocol (KrinosAI lifecycle pack, 05-References/). Items are labelled Observed (code or Git at a6b4429), Measured (dated run), Stated (named person, dated) or Not recorded. Test records describe the tests that exist at a6b4429 and the one run made on 2026-10-28. No earlier test runs, test plans or sign-offs were recorded; tests for untested requirements are planned, not run. Dataset history is held, if anywhere, in the external Visionary_oral_ML repository, which is not part of this baseline.
Acceptable evidence: an implementation reference (file:line at a commit SHA), a dated test record, a signed decision, a vendor agreement, or another independently verifiable artifact. Never invent approvals, clinical results, measurements or performance data.
1. Dataset version source and license
| Dataset |
Version |
Source |
Rights basis |
| Training data of cavity, calculus, gum_recession, swollen_gums, teeth_seg models |
Not recorded — OPEN-044 |
Not recorded — models trained in Visionary_oral_ML (external) |
Not recorded — OPEN-010, OPEN-044 |
Export check images mobile/ml/sample-images/ (front, upper, lower) |
At a6b4429 |
Repository |
Not recorded — OPEN-044 |
2. Collection time location and devices
| Dataset |
Period |
Sites |
Devices |
| Training data |
Not recorded — OPEN-044 |
Not recorded — OPEN-044 |
Not recorded — OPEN-044 |
3. Patient cohort inclusion and exclusions
| Criterion |
Rule |
Count affected |
| Patient cohort (age, gender, dentition, clinic) |
Not recorded — OPEN-044 |
Not recorded — OPEN-044 |
4. Labels annotation protocol and QA
| Item |
Entry |
| Label classes (Observed) |
cavity, calculus, gum_recession, swollen_Gums, teeth (mobile/ml/model-manifest.json) |
| Annotation protocol, annotators and QA |
Not recorded — OPEN-044 |
5. Split strategy leakage and duplicates
| Split |
Images |
Method |
| Train / validation / test |
Not recorded — OPEN-044 |
Not recorded; no leakage or duplicate check recorded — OPEN-044 |
6. PHI deidentification consent and retention
| Item |
Entry |
| De-identification |
Not recorded — OPEN-044 |
| Consent and legal basis for training use |
Not recorded — OPEN-009, OPEN-031 |
| Retention |
Not recorded — OPEN-029 |
7. Bias coverage limitations and release record
| Limitation |
Effect |
| No documented dataset |
Performance cannot be claimed for any population; subgroup bias (age, skin tone, phone camera, lighting) is unknown |
| Single-class detectors with a low confidence threshold (0.1) |
More false positives expected; validation must measure them (KAI-VORAL-VAL-001) |
Open issues and decisions
| ID |
Specific gap or decision |
Responsible owner |
Due date |
Blocks gate? |
Evidence required |
Status |
| OPEN-001 |
Tracker project ID, named requester and sponsor, and named reviewers are not recorded |
Product Owner |
2026-11-10 (proposed) |
Yes |
Tracker entry; names with roles |
Open |
| OPEN-009 |
No DPIA |
Security & Privacy Officer |
2026-11-30 (proposed) |
Yes |
DPIA |
Open |
| OPEN-010 |
Model and dataset rights |
ML Lead |
2026-11-30 (proposed) |
Yes |
Rights register |
Open |
| OPEN-029 |
No retention or deletion periods for patient data, photos, reports, backups or phone caches |
Security & Privacy Officer |
2026-11-30 (proposed) |
Yes |
Retention schedule in the DPIA |
Open |
| OPEN-031 |
No recorded patient consent or privacy notice for photo capture, AI processing and QR self-registration |
Security & Privacy Officer |
2026-11-30 (proposed) |
Yes |
Consent text, where it is shown, and stored consent records |
Open |
| OPEN-044 |
Training and evaluation datasets of the five on-device models (source, rights, consent, cohort, labelling protocol, splits) are not recorded in this repository; the models were trained in the external Visionary_oral_ML repository |
ML Lead |
2026-12-15 (proposed) |
Yes |
Completed dataset datasheet with source and rights basis |
Open |
Review and approval
| Review role |
Named individual |
Decision |
Date |
Signed / controlled record |
| Author |
Claude Code (drafted with Claude Code) |
Draft |
2026-10-28 |
This file at the commit that adds it |
| Technical reviewer |
Tech Lead (name to be assigned) |
Pending |
Pending |
Pending |
| Quality / regulatory reviewer (if applicable) |
Qualified regulatory reviewer (name to be assigned) |
Pending |
Pending |
Pending |
| Authorized approver |
ML Lead (name to be assigned) |
Pending |
Pending |
Pending |