The publishing standards

Editorial Policy: Evidence, AI Use and Corrections

A useful explanation should let you trace a claim to its source. These standards describe how this site handles research, AI assistance, illustrations, measurement claims and corrections, including the distinction between source checking and independent expert review.

01 · The evidence boundary

Read the claim, its support and its limit together

Three different statements need three different kinds of support. The distinction matters when a precise-looking number feels more authoritative than the evidence behind it.

Product behavior

“The formula returns 96.”

An invented mean mismatch of 0.04 gives 100 − 100 × 0.04 = 96. Code inspection and arithmetic can check that transformation.

Scientific finding

“A study observed a relationship.”

Read who was studied, what was measured and how. A finding about edited portraits does not automatically describe every natural face.

Unestablished inference

“96 means healthy or attractive.”

That conclusion does not follow from this tool’s equation. We have no validated health or attractiveness interpretation for its index.

For a fictional mismatch of 0.04, the defined equation gives index 96. The calculation alone supplies no evidence of health or attractiveness.
Original teaching diagram with invented input. It illustrates a reasoning boundary, not a participant result or validation study.
02 · Who and how

AI assistance is part of the process

The publisher is Face Symmetry Checker. AI coding assistants help find sources, draft and revise text, create diagrams, inspect implementation and run software checks. That work can contain mistakes; fluent language and automated checks are not expert endorsement.

We do not substitute invented names, professional credentials, testimonials or generated author portraits for accountable authorship. There is currently no named independent medical reviewer. When a page says “research checked,” read it as a source-and-content check, not medical peer review. About this publisher gives the available background.

Our use of AI disclosure follows the transparency principle described in Google’s people-first content guidance. That guidance is not an approval of this site.

03 · Choosing support

Use sources that answer the particular question

Technical behavior
Use the actual implementation and the library’s official documentation. MediaPipe documents estimated landmarks; our separate scoring convention is documented in the methodology.
Research findings
Prefer original studies and relevant evidence reviews. State the measurement, sample and limitations when they affect interpretation. Preserve uncertainty and disagreement rather than turning an association into a cause.
Health information
Use named public-health agencies and professional clinical sources for general guidance. Refer questions about a person’s symptoms to appropriate care. A photo score cannot select a diagnosis or treatment.
Products and reader questions
Attribute vendor claims to the vendor and distinguish desk research from hands-on testing. Forums and social posts can reveal questions to address; popularity and personal anecdotes are not validation.

For health information, also check the publisher, purpose, sources and review date. The National Library of Medicine’s evaluation guide explains those checks. Citing it does not imply institutional endorsement.

04 · Numbers and pictures

Keep measurements and illustrations in context

The main checker compares 16 selected landmark pairs in a two-dimensional image. Its displayed index is a product convention, not a percentage of a face that is symmetrical, a population percentile or a clinical threshold. The score guide explains the calculation and rounding.

Equation tests check expected behavior for defined inputs. Browser checks exercise controls, calculations and exports. Neither establishes anatomical accuracy, demographic fairness or performance on every physical device. Our validation status states the work still missing. Adding points or weights is not, by itself, evidence of greater accuracy.

Original diagrams explain geometry and are labeled when they use invented values. The built-in portrait is a fictional AI-generated example; its displayed analysis comes from the tool. Mirror composites repeat one half of an image. These are demonstrations, not patient photographs, treatment outcomes or evidence of a better-looking side.

Photo handling is a separate question from publishing quality. The current tools process selected photos and measurements locally in the browser. Loading the website still makes ordinary hosting requests. Downloads, extensions and browser agents you enable can create or access copies; supported browser agents can read the current numerical report. Read the privacy information before choosing a personal photo.

05 · Review and change

A date needs a specific meaning

A page’s source-check date identifies its stated content review. It is not the publication date of each cited study, a promise of continuous monitoring or proof of clinician review. A tool’s exported method version identifies its calculation; it does not establish a clinical certification.

Our standard is to revisit a page when a relevant source, product or calculation changes, or a substantiated error is reported. Substantive corrections should explain what changed on the affected page. We do not advance a review date merely because a file was rebuilt or a link label changed.

This revision: October 27, 2026 - added explicit AI-use disclosure, evidence boundaries, review-date definitions and correction instructions.

06 · Commercial context

Make the publisher’s interest visible

The current photo tools are free and require no account or checkout. Our app comparison includes our own tool and discloses that interest. That guide states there are no affiliate links or paid placements. It reviews public documents; it is not a third-party benchmark.

Our standard is to label a sponsored placement or affiliate relationship beside the relevant recommendation if one is introduced. Free access alone does not establish independence or accuracy, and this policy is not a funding audit.

07 · Corrections

Help us locate an error without sending a face photo

Email contact@face-symmetry-checker.com with the page URL, the exact statement or control, why it appears wrong, and a supporting source or reproducible steps. For a tool issue, add the browser version and exported method version if available. Remove personal measurements or identifiers you do not want to share.

A description or a fictional example is usually a better starting point than a personal portrait. This mailbox is for site feedback, not diagnosis or urgent care. See contact information for other inquiry types.

Use the policy

Before relying on a claim

Use these four checks on a page here or elsewhere. There is no trust score or submission.

Can you identify each of these?

Checkboxes use the current page only. This checklist does not certify the source.

Questions about these standards

Does an official citation validate this checker?

No. A source can explain a detector or report a specific study without evaluating our implementation or score. Check whether the evidence addresses the exact claim.

Is every article independently medically reviewed?

No independent clinician review is currently claimed. The guides provide educational context and link to sources; they do not replace assessment by a qualified professional.

Can I report a problem without uploading a photograph?

Yes. Send the page URL, a description and reproduction steps or a supporting citation to the published contact address. Avoid personal images and unnecessary identifying details.

Publishing responsibility

Face Symmetry Checker publishes these pages. Aron is the public site contact for questions and corrections. AI assistance, source checking and software verification are disclosed separately from independent professional review. A source-check date does not imply a named clinician approved the page.

Privacy settings

Photo analysis works without analytics. You can optionally allow Google Analytics to measure basic page visits. We do not send photos, filenames, face measurements or scores to analytics.

Analytics is off unless you allow it.

Your choice is stored in this browser. Turning analytics off after allowing it reloads this page; save any wanted result first. Read the privacy policy.