Simulation With a Bad Photo: What to Do Before Showing It

The simulation is only as good as the photo it comes from. A well-made frontal image allows a coherent projection; a crooked, poorly lit photo, or one with the patient smiling, can produce an image that misleads, the worst scenario in a consultation meant to align expectations. This page shows the signs that a photo is not good enough, what to redo, and how to explain it to the patient without embarrassment.

Why is the photo so decisive?

Because the projection is generated from the photo, and the AI preserves what is in it: light, angle, expression, background. If the face is tilted, the "after" will be too. If there is a strong shadow on one side, the change can look like asymmetry. If the patient smiles, the mouth area is read in a different position. The result can look convincing and still be wrong. See standardized clinical photography and standardized frontal photo.

Which signs show the photo is not good enough?

If any item fails, retake it. The most common are the smile, an angle slightly from above or below, and office light coming from one side. See also white balance and focal length, which affect the fidelity of color and proportion.

How do you retake the photo without embarrassing the patient?

Speak simply and without blaming. "I need a more standardized photo so the image is faithful." Position the patient, adjust the light, ask for the neutral expression, and check before generating. Two minutes spent here save an entire consultation of noise. See how to use the frontal photo in the conversation.

What if you cannot retake it?

If the patient cannot or does not want to retake it, explain the limitation. You can postpone the simulation to a follow-up, lead the consultation with a mirror and an area map, or generate the image for internal use only, without showing it. Record the decision. See when not to show the simulation and how to use the mirror in the consultation.

What does Harmy require from the photo?

The app works with a single frontal photo per session, which can come from the camera or the gallery. The face is detected on the device itself, and the marking points depend on that reading. The cleaner the photo, the better the marking and the projection.

What should you record?

Record whether the photo was taken in the office, the conditions, and whether the patient authorized sending it to the third-party AI. If you retook the photo because of a quality problem, note it. That helps justify why you did not show the first image. See patient image security in the simulation.

A standardized photo is the cheapest investment for a reliable simulation. If the image is to be a good illustration, the photo has to be a good starting point.

How do you train the team to photograph well?

Photo quality is a process, not a talent. Make a one-page guide with the patient's position, the distance from the camera, the height of the device, the light, and the expression. Train whoever takes the photos and check the first photos of each new team member. A neutral background, a mark on the floor, and a fixed light make the process repeatable.

Review the guide every six months or when the space changes. When everyone photographs the same way, before and after comparisons are fairer and the simulation more reliable. See standardized clinical photography.

With a bad photo, own the limit: ask for a new image, straight on and in good light, before showing the patient any projection. If the image still does not improve, say that the simulation turned out less reliable and note it. None of this replaces clinical assessment, which is individual.