Common mistakes that can distort a face comparison

Many confusing results come from the photos rather than the people being compared. These are the most common problems to check before judging a similarity percentage.

Using a face that is too small in the image

Group photos and distant full-body images may leave too little facial detail. Crop or choose a portrait where the face is clearly visible rather than forcing the model to work from a tiny region.

Mixing a frontal photo with a strong side profile

Comparing very different head angles changes which features are visible and how their spacing appears. Near-frontal images are usually easier to compare consistently.

Using beauty filters on only one image

Filters can enlarge eyes, smooth skin, reshape the jaw or nose and alter proportions. If only one photo is filtered, the model may be comparing editing effects as well as natural facial traits.

Ignoring large age differences

A young child and an adult parent naturally have different proportions. Even photos of the same person taken many years apart can produce different results. Age should be considered when interpreting the percentage.

Choosing heavily compressed screenshots

Messaging apps, social networks and repeated screenshots can reduce detail. Whenever possible, use the original or a reasonably high-quality image rather than a heavily compressed copy.

Treating one number as a factual conclusion

The most important mistake is giving the score more meaning than it has. It is an estimate of visible resemblance under particular image conditions. It is not a genetic measurement or a family relationship test.