CHEEK QUESTION / 1 NOVEMBER 2026
Cheek asymmetry in photos: does this test measure it?
If one cheek looks fuller in a photograph, this test cannot quantify that fullness difference or explain its physical cause. It returns an overall index and Eyes, Eyebrows, Nose, Mouth and Jaw fields, with no separate Cheek score. Do not relabel the total or Jaw output as a cheek measurement. Our saved records show why those distinctions matter.
Watch what this tool can and cannot return
English transcript
This photo tool does not return a separate cheek symmetry score. If one cheek looks fuller, the overall result cannot quantify that difference. Read the field names before using a number as an answer. Our saved records contain an overall index and five fields. They are eyes, eyebrows, nose, mouth and jaw. No independent cheek measurement was recorded. That does not mean every detected point avoids the cheek area. It means a cheek result is not among the saved outputs. In the October twenty sixth file test, the original total was 98.0 and the jaw field was 98.6. Flat rotation raised the total to 98.1 while lowering jaw to 98.4. The JPEG copy returned 97.5 overall and 96.9 for jaw. Each file was uploaded three times. Renaming any of those numbers would not make it a cheek measurement. Ten unchanged-file uploads repeated all displayed fields. Their displayed ranges were 0.0, but that is not a cheek score or proof of accuracy. The separate camera case included six turned-input refusals. Every refusal had no total and no feature values. Keep the original, inspect the overlay, and record the labels and refusals. For cheek fullness, our records do not establish a physical difference or its cause. See the full dated records on our cheek asymmetry page. Photo geometry only, not a beauty rating or medical assessment.
Find the measurement before interpreting the number
Check what the results panel actually calls each field. The site’s current output labels were checked on 1 November 2026: Eyes, Eyebrows, Nose, Mouth and Jaw, plus the overall index. There is no dedicated Cheek output in that panel or the saved records below. This does not mean no detected point lies near a cheek; it means we did not record an independent cheek measurement.
We have no saved cheek-volume measurement, physical reference scale or signed left-minus-right cheek distance in these observations. We do not know which cheek is physically different or why from this dataset. These limits apply to our records, not a claim about every analysis tool.
One input can move the fields in different directions
The 26 October 2026 prepared-file case used one public portrait, five files and three uploads per file. The original returned overall98.0, eyes98.0, eyebrows98.2, nose98.2, mouth97.1 and jaw98.6 on each of three uploads. None of those labels was Cheek.
Flat +10° image rotation returned overall98.1, eyes99.1, eyebrows98.9, nose97.8, mouth96.2 and jaw98.4 on every upload. Compared with the original, the total increased by 98.1−98.0=0.1 points while Jaw decreased by 98.4−98.6=−0.2. Nose decreased by 0.4 while Eyes increased by 1.1. There is no defensible way to choose one of these values and rename it a cheek result.
The JPEG quality20 copy with subsampling2 returned overall97.5 and jaw96.9 on all three uploads. Their differences from the original were −0.5 and −1.7 points. Gamma1.4 and LANCZOS256 copies remain in the full table too. These were prepared files of one portrait, not measurements of cheek tissue changing.
Matching repeats still do not measure a cheek
A separate ten-upload check of the unchanged file on 26 October repeated the same six fields: overall98.0, eyes98.0, eyebrows98.2, nose98.2, mouth97.1 and jaw98.6. Every displayed field had maximum minus minimum0.0. Repeatability at that precision does not supply a missing Cheek field or establish physical accuracy.
The camera case, recorded 21 October, used another person’s four photographs with three attempts each. Near/front returned overall98.6 and jaw99.1 three times; far/front returned overall96.1 and jaw92.5 three times. Distance was not isolated, and these were not cheek-fullness measurements.
Both turned photographs were refused three times each. All six attempts returned no total and no feature scores, with the exact message “Your head appears turned. Use a photo facing straight toward the camera.” Those attempts do not produce a cheek score, a low score or a value to fill in later.
Use the photograph without inventing a cheek verdict
Open the face symmetry test if you want to inspect its detected overlay. Keep the original image, read the actual field names and retain the input conditions. If your question is specifically cheek fullness, record that this tool has not measured it rather than selecting the closest-looking number.
The photo variables guide covers the wider input comparisons. Our nose-field page answers a different question: how an output that actually exists can diverge from the total. This tool measures photo geometry, not beauty, and is not a medical assessment.
Three questions about cheek asymmetry and this tool
Is Jaw another name for Cheek?
No. Jaw is the label used by our scoring groups and recorded output. These observations provide no validated conversion from it to a cheek-fullness measurement.
Does a high overall result mean both cheeks match?
That conclusion is not supported by these records. No independently calibrated cheek comparison was collected.
Should the missing Cheek value be written as zero?
No. Unmeasured is not zero. The displayed0.0 in the unchanged-file check is a repeat range for existing fields, not a cheek score.
Every original record, including refused inputs
These are the original 37 attempts: 12 camera attempts including six refusals, 15 prepared-file uploads and ten unchanged-file repeats. The measurement dates stay 21 and 26 October 2026. There were no new captures or analysis runs for this article.
Show the dates, totals, five feature fields, repeats and refusal messages
| Date | Input / conditions | Repeat | Overall | Eyes | Eyebrows | Nose | Mouth | Jaw | Recorded status |
|---|---|---|---|---|---|---|---|---|---|
| 2026-10-21 | 150 Far / turned; About 0.8 m / about 15–20° | 1 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 150 Far / turned; About 0.8 m / about 15–20° | 2 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 150 Far / turned; About 0.8 m / about 15–20° | 3 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 151 Far / front; About 0.8 m / front-on | 1 | 96.1 | 95.5 | 97.5 | 96.6 | 98.1 | 92.5 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-21 | 151 Far / front; About 0.8 m / front-on | 2 | 96.1 | 95.5 | 97.5 | 96.6 | 98.1 | 92.5 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-21 | 151 Far / front; About 0.8 m / front-on | 3 | 96.1 | 95.5 | 97.5 | 96.6 | 98.1 | 92.5 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-21 | 152 Near / turned; About 30–40 cm / about 15–20° | 1 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 152 Near / turned; About 30–40 cm / about 15–20° | 2 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 152 Near / turned; About 30–40 cm / about 15–20° | 3 | — | — | — | — | — | — | Your head appears turned. Use a photo facing straight toward the camera. |
| 2026-10-21 | 153 Near / front; About 30–40 cm / front-on | 1 | 98.6 | 99.0 | 98.5 | 98.6 | 98.0 | 99.1 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-21 | 153 Near / front; About 30–40 cm / front-on | 2 | 98.6 | 99.0 | 98.5 | 98.6 | 98.0 | 99.1 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-21 | 153 Near / front; About 30–40 cm / front-on | 3 | 98.6 | 99.0 | 98.5 | 98.6 | 98.0 | 99.1 | Analysis complete. Try another photo under the same conditions. |
| Date | Input / conditions | Repeat | Overall | Eyes | Eyebrows | Nose | Mouth | Jaw | Recorded status |
|---|---|---|---|---|---|---|---|---|---|
| 2026-10-26 | Unchanged PNG512 | 1 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Unchanged PNG512 | 2 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Unchanged PNG512 | 3 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Gamma1.4 darkening | 1 | 97.6 | 98.1 | 97.8 | 98.1 | 96.2 | 98.1 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Gamma1.4 darkening | 2 | 97.6 | 98.1 | 97.8 | 98.1 | 96.2 | 98.1 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Gamma1.4 darkening | 3 | 97.6 | 98.1 | 97.8 | 98.1 | 96.2 | 98.1 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | +10° flat rotation | 1 | 98.1 | 99.1 | 98.9 | 97.8 | 96.2 | 98.4 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | +10° flat rotation | 2 | 98.1 | 99.1 | 98.9 | 97.8 | 96.2 | 98.4 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | +10° flat rotation | 3 | 98.1 | 99.1 | 98.9 | 97.8 | 96.2 | 98.4 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | LANCZOS256 px PNG | 1 | 98.1 | 98.5 | 98.0 | 98.9 | 97.7 | 97.7 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | LANCZOS256 px PNG | 2 | 98.1 | 98.5 | 98.0 | 98.9 | 97.7 | 97.7 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | LANCZOS256 px PNG | 3 | 98.1 | 98.5 | 98.0 | 98.9 | 97.7 | 97.7 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | JPEG quality20 subsampling2 | 1 | 97.5 | 98.6 | 98.7 | 97.1 | 96.4 | 96.9 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | JPEG quality20 subsampling2 | 2 | 97.5 | 98.6 | 98.7 | 97.1 | 96.4 | 96.9 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | JPEG quality20 subsampling2 | 3 | 97.5 | 98.6 | 98.7 | 97.1 | 96.4 | 96.9 | Analysis complete. Try another photo under the same conditions. |
| Date | Input / conditions | Repeat | Overall | Eyes | Eyebrows | Nose | Mouth | Jaw | Recorded status |
|---|---|---|---|---|---|---|---|---|---|
| 2026-10-26 | Exact unchanged baseline file | 1 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 2 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 3 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 4 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 5 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 6 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 7 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 8 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 9 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
| 2026-10-26 | Exact unchanged baseline file | 10 | 98.0 | 98.0 | 98.2 | 98.2 | 97.1 | 98.6 | Analysis complete. Try another photo under the same conditions. |
Read the original combined reference copy. The camera participant and public portrait remain separate datasets.