Face Occlusion
Face Occlusion is a function to detect occluded objects to support verification business cases. The face occlusion algorithm is trained using facial features. The algorithm will check the face area on the forehead, right eyebrow, left eyebrow, right eye, left eye, right cheek, left cheek, nose, and mouth. If one of the facial features is occluded, the liveness will reject the selfie.

Facial attributes consist of forehead, eyebrows, eyes, cheeks, nose, and mouth.
The face occlusion algorithm is combined in our liveness check, it will be performed before the general liveness check. By this feature, we help you ensure the selfie image is visible with no occluded object on the selfie photo so that your verification process will be unbiased to identify your customer. Here the steps on our liveness check are enhanced by the occlusion check function.

Liveness Check Flow
{
"job": {
"id": "4b4441eead2921862ddceebdce205b44ff50aac8960f5102GIYDEMJNGEZC2MBZ",
"result": {
"status": "incompleted",
"analytic_type": "FACE_LIVENESS",
"result": []
}
},
"message": "the face is most likely occluded",
"ok": true
}
The occluded objects, such as masks, sunglasses will be rejected by the algorithm.


Case | Accepted | Rejected |
---|---|---|
Woman with hijab | ![]() | ![]() |
Eyeglassess | ![]() | ![]() |
Hair on forehead | ![]() | ![]() |
Another object occlusion | ![]() | ![]() |
Case | Example |
---|---|
Facial Paint or tattoo | ![]() |
Mask with face pattern | ![]() |
Last modified 10mo ago