Liveness-Certel
Liveness detection is the ability of a system to detect if a face is real or fake. Here are some examples of performing liveness:
- Only one face exists in the image sent
- The face in the image must not be blocked by anything, i.e., take off masks, sunglasses/light reflection on glasses, hat
- Make sure the image has good quality (avoid blur & dark)
- The minimum resolution of the image is Height: 640 x Width: 480
- Minimum size for each image: it’s recommended to be 200kb with minimum quality is JPEG 90%
- Avoid backlight
- Avoid taking images in a dark environment
- No editing on the image
- prioritize for users to directly capture face images from mobile phone camera
- do not compress or crop the image
- do not add any elements to the image, i.e., padding
Case examples:
- Look straight (face should be looking straight)
- Lighting condition (avoid taking images in backlight or low light conditions. It may reduce the visibility of the facial features)
- Orientation (the orientation of the face should be portrait with the correct orientation)
- Face in the center (the face should be in the center to avoid half-face images which might lead to no face detected)
- Background condition (avoid taking the image in a crowded background as it might affect the result)
- No attributes covering the face (avoid wearing attributes or anything that cover your faces such as sunglasses/glasses with light reflection, face masks, hat, and hair)
- No blur (the blurrier the face photo, the less liveness accuracy)
- Not holding any identity card.
Updated 10 months ago