How to Take a Photo for an AI Treatment Simulator

A treatment simulator can only work with what the source photo actually shows. Soft frontal light, a level camera, a relaxed expression, full treatment-area visibility, and no filters make it easier to create a bounded illustration and easier to notice when the result changes something it should not.

1. Face soft, even light

Stand facing a window or another broad, neutral light source. The goal is for both sides of the face and the treatment area to be visible without a bright hotspot or deep shadow. Strong overhead light can exaggerate under-eye and submental shadows. Backlight can make the face too dark. Colored bulbs can shift skin tone.

Do not chase flattering light at the expense of consistent light. A dramatic portrait may look attractive but give the simulator less stable information to work with.

2. Keep the camera at face height

Hold or place the phone approximately level with the center of the face and far enough away to avoid a wide-angle close-up. Looking up from a low camera changes the apparent jaw and neck. Looking down toward a high camera can change the forehead, eyes, cheeks, and lower-face proportions.

Keep the phone upright, the head level, and the face centered. Use the normal camera rather than an in-app beauty camera. If the phone offers automatic lens correction or face retouching, turn retouching off.

3. Use a straight-on pose and relaxed expression

Start with a front-facing photo unless the product explicitly asks for another view. Keep the chin neutral rather than lifted or tucked. Close the lips gently and relax the brow, eyes, jaw, and cheeks.

A wide smile changes the visible relationship among lips, cheeks, eyes, nasolabial folds, jaw, and teeth. A raised brow changes forehead lines. Those expressions may be useful in a provider assessment, but they make a simple before-and-after comparison harder unless the workflow requests them deliberately.

4. Include the full face and target area

Do not crop through the forehead, chin, jaw, or neck when those regions could matter. Leave a small margin around the head. Make sure the selected area is large enough and sharp enough to inspect.

• Forehead or expression-line preview: include the full forehead and both brows.
• Lip or cheek preview: keep lips relaxed and include both sides of the face.
• Under-eye preview: remove glasses and avoid overhead shadow.
• Jawline or submental preview: show the full jaw, chin, and upper neck with hair and collars moved away.
• Skin-tone or texture preview: use neutral light and a sharp source without makeup filters.

5. Move obstructions away

Hair across the cheek or jaw, fingers touching the face, large frames, masks, hats, high collars, and strong jewelry reflections can hide an edge the system needs to preserve. Move them only enough to reveal the area; do not restyle the entire portrait between comparisons.

6. Turn off filters, portrait retouching, and strong blur

Beauty filters may narrow the face, enlarge eyes, smooth texture, recolor lips, whiten teeth, or change lighting before generation begins. Portrait mode can blur hair, ears, and jaw edges. Social screenshots may add compression, overlays, or prior edits.

Use a recent image from the normal camera and inspect it before uploading. If the source already contains a face-changing edit, the simulation cannot provide a trustworthy comparison to an unedited appearance.

7. Check focus, exposure, and motion

Zoom in briefly. The eyes, lips, skin texture, and target-area edges should be in focus. Retake motion blur, heavy compression, blown highlights, very dark shadows, lens smudges, and glare. Cleaning the phone lens often fixes the soft haze that looks like missed focus.

When the right answer is to retake the photo

Retake instead of forcing a generation when the image is a profile, the face is cropped, more than one face is present, the target is obstructed, lighting hides one side, a filter is visible, or the file is too soft to inspect. A clear rejection is safer and more useful than a polished simulation built on an ambiguous source.

Inspect the result against the source

After generation, compare more than the treatment area. Check hair, brows, eyes, nose, teeth, ears, jewelry, skin tone, lighting, face shape, clothing, and background. If unrelated features changed materially, do not treat the image as a good reference just because the intended edit looks attractive.

The simulation remains an illustration even when the source is excellent. It cannot evaluate anatomy, diagnose a concern, select a treatment, or predict an individual's result. Bring the visual preference to a licensed provider.

You can apply these steps in Eva's free AI before-and-after preview or begin with a treatment-specific route such as the lip filler simulator or jawline simulator. For a deeper quality review, use the simulation evaluation framework.