OPTICAL LAB / 001 · LOOK CLOSER

Your glasses.
A different perspective.

Crooked frame or camera trick? Reflect, align, and investigate what your photo is really showing.

ORIGINAL / REFLECTION / PERSPECTIVE
PHOTO STAYS ON YOUR DEVICENO SIGN-UPMANUAL, TRANSPARENT OBSERVATIONS
GHOST MIRROR ENGINE
SAMPLE 01 / ILLUSTRATIONAXIS 50.0%
Illustration, not a physical diagnosis
OriginalReflectionAxis
JPEG / PNG / WebP · 15 MB

Front-view frame photo. Place the axis through the bridge. Sample outlines are illustrative.

Sample loaded. Uploads stay in this browser. No automatic frame detection.
01 / CAPTUREStart with a clear view

Keep the camera level and the frame centered.

02 / REFLECTFollow the ghost

Move the bridge axis, then compare mirrored details.

03 / INVESTIGATEQuestion the difference

Repeat the photo before interpreting a mismatch.

01 / OBSERVATION RECORD

Evidence, before conclusions.

These numbers describe your manually placed points in the 1000 × 570 comparison canvas, not physical frame dimensions.

MANUAL POINTS 4 / 4
PHYSICAL CAUSE UNDETERMINED

Illustrative sample observation

—Lens-center vertical difference · px
—Axis-distance difference · px

No normal range, medical score, millimeter measurement, or lens optical-center assessment is inferred. Reflections and facial asymmetry can affect appearances.

02 / PHOTO TRUTH LAB

Make the photo tell a fairer story.

A photograph flattens a three-dimensional object. An uneven outline is a clue, never proof of a bent frame.

A / FRAME ANATOMY

Frame without face

Photograph the frame straight on against a plain contrasting background. Keep both hinges equally visible. Avoid pressing the frame into a symmetrical shape.

B / WEARING BALANCE

Fit is a different question

Keep your head relaxed and look straight toward a level camera. Compare frame position relative to facial landmarks. Natural differences in ears, nose, and face can matter.

C / MULTI-VIEW

One view is never the whole frame

Repeat the front photo, then add top and side views. Compare like with like. A side view and a front view cannot be overlaid as a deformation measurement.

03 / GHOST TIMELINE

Keep a trail of observations.

Explicit saves only. Up to six comparison snapshots stay in this tab until it is closed or reloaded. Download anything you want to keep.

Your session starts here. Save a view to compare it later.
04 / OPTICAL ACADEMY

Learn to look, not just measure.

Perspective playground

Move the illustrative camera angle. One side appears narrower even though this conceptual frame has no physical damage. This is a diagram, not a camera calibration model.

The crooked glasses case

A single close-up photo shows the right lens larger than the left. What can you conclude?

Choose a conclusion to reveal the reasoning.

A mismatch is a question.

Repeat the view. Check the camera. Follow the evidence.

Back to the experiment ↗
READ / 01

Why glasses may look tilted

Frame position, head tilt, camera roll, perspective, and facial asymmetry can overlap. Start by leveling the camera, then repeat the photo. A visible difference alone does not identify the cause.

READ / 02

Transparent and rimless frames

Glare, thin edges, and lens reflections make outlines ambiguous. Use a contrasting background and soft even light. Place anchors on repeatable structural points instead of bright reflections.

READ / 03

Before and after fitting

Keep distance, camera height, orientation, and frame position consistent. Overlay photos to record appearance, not to infer optical quality. An optician can examine support, stability, and the frame in person.

05 / QUESTIONS, IN FOCUS

A little more clarity.

Can this prove my glasses are bent?

No. It compares a two-dimensional photograph and user-selected points. Three-dimensional twist, lens optical centers, prescription accuracy, and visual effects require other methods or in-person assessment.

Are photos uploaded to a server?

The workbench reads selected photos locally and does not send them to an analysis server. Session snapshots remain in memory. Blogger and Google advertising may use their own cookies and requests independently of photo processing.

What does pixel difference show?

It compares color values in the original and reflected images. Lighting, skin, background, and glare also produce differences. It is not an automatic frame-edge detector or defect map.

Should I bend the frame myself?

This tool provides no physical adjustment instructions. For discomfort, loose or damaged parts, or persistent uneven fit, take the frame and your observations to an optician.

What is Frame DNA?

An SVG of the four manually placed anchors and reflected guide. It contains no source photo. It is a minimal geometric sketch, not an automated identity or biometric fingerprint.

ABOUT / PRIVACY / LIMITS

A small studio for a closer look.

FRAMEGHOST is an independent visual comparison tool. There is no automatic diagnosis, professional certification, or affiliation with an eyewear brand. Source images are not saved automatically. Use Try sample to return to the sample; reload to clear session snapshots. Downloaded images may contain your photo: share them thoughtfully.