Face Not Matching
Understanding and fixing this common KYC rejection
⚠️ What This Error Means
This error fires when the algorithm comparing your live selfie to your ID photo calculates a similarity score below the platform's approval threshold. It is not a yes or no judgment on whether you are really you, it is a numeric confidence score, and something about the two images pushed that number too low.
Facial matching systems do not compare identities, they compare pixels and geometry. Research on face recognition accuracy consistently points to four factors that degrade a match: pose, illumination, expression, and aging. Every one of those can be present in a completely legitimate selfie from a completely real person. The system has to accept normal variation between two photos of the same face while still rejecting a different face entirely, and that balancing act is exactly why honest users get bounced.
There is no single public number for how strict that threshold is. Vendors set and retune it per platform based on their own risk tolerance, and none of them publish it, since doing so would just teach fraud attempts where the line sits. What is documented is which inputs move you further from a match: a longer gap in years since your ID photo was taken, a non frontal angle, occlusion from glasses or facial hair, cosmetics that change how the descriptor reads your face, and uneven or backlit lighting. None of these mean the system thinks you are lying. They mean the two images gave it less to work with than it wanted.
✅ How to Fix It
- Remove glasses, hats, and heavy makeup. Take off anything covering your hairline, brows, or eyes, and skip heavy makeup that changes your facial contours from how your ID photo shows you.
- Use even, direct lighting. Face a window or a light source straight on so light falls evenly across your whole face. Avoid shadows across one side, under your chin, or behind you.
- Look directly at the camera, straight on. Hold the phone at eye level and face it head on rather than tilting your head or shooting from above or below.
- Match your expression to your ID photo. Keep a neutral, closed mouth expression close to how you look in your ID rather than smiling or pulling a face.
- Use a current, unedited photo of your actual face. Skip filters and beauty modes entirely, and if your ID photo is old, lean harder on getting everything else right to close the gap.
What Actually Hurts Your Match Score
| Factor | Why It Matters | What To Do |
|---|---|---|
| Outdated ID photo | Elapsed time between the reference photo and your selfie is a documented factor in match accuracy. Aging gradually changes facial structure, skin texture, and weight distribution in ways the algorithm has to account for. | You cannot change the ID photo, so compensate by making the selfie as clean and unambiguous as possible. Sharp focus, even light, no expression games. |
| Glasses or heavy makeup | Glasses occlude the eye and brow region and often add glare. Heavy cosmetics change the facial descriptor the algorithm extracts, especially around contouring and eyebrow shape. | Remove glasses unless your ID photo also shows glasses. Keep makeup close to natural, skip heavy contour for this one photo. |
| Poor lighting or shadows | Uneven, backlit, or low light is one of the most cited causes of reduced matching accuracy, since it distorts how the system reads contours and skin tone. | Face a window or lamp directly, never have the light source behind you. Avoid dim rooms and overhead-only lighting. |
| Extreme angle | Pose is one of the four classic factors that hurts recognition accuracy. Most reference databases and algorithms are tuned around frontal faces, not tilted or profile angles. | Hold the camera at eye level, chin level, straight ahead. Do not shoot from above, below, or a side angle. |
| Facial hair changes | A new beard, a shaved beard, or a different mustache shape than your ID photo changes jaw and cheek geometry and can occlude the lower face entirely. | If you can match your ID photo's facial hair, do it for this attempt. If not, make sure lighting and angle are otherwise perfect to offset it. |
Let Us Help You Fix This
Try our free AI analysis first, or let our experts fix it for you.
Frequently Asked Questions
Why does my selfie not match my ID if it is really me?
Facial matching compares pixels and geometry, not identity, so real changes like aging, weight change, lighting, or glasses can drop your similarity score below the system's threshold even though you are genuinely the person in the ID. It is a false rejection of a real match, not evidence of anything wrong.
Does an old ID photo cause more rejections?
Yes. The number of years between the two photos is a documented factor in facial matching accuracy, since aging gradually changes facial structure and skin texture. The older your ID photo, the harder the system has to work to confirm a match, so your selfie needs to be as clean and well lit as possible to compensate.
Should I remove my glasses for the selfie?
Generally yes, unless your ID photo also shows you wearing glasses. Glasses can create glare and partially occlude the eye and brow area the algorithm relies on, and mismatched glasses between the two photos is one of the more common causes of a low match score.
How many times can I retry the face match before something worse happens?
Most platforms allow a handful of attempts before locking the account or forcing manual review, though the exact number is rarely published. If you have failed twice with clear, well lit, unedited photos, stop guessing and contact support, since repeated rapid retries can themselves trigger a separate fraud flag.
Can lighting alone cause a face match failure?
Yes. Uneven, backlit, or shadowed lighting is one of the most common and most fixable causes of a low match score, since it changes how the algorithm reads your facial contours and skin tone. Facing a window or a soft light source directly usually resolves this on its own.