
Technical Details of Ayonix Face-ID SDK
Ayonix Face-ID SDK finds human faces from image and recognizes/authenticates if there is an appropriate record of the person. Ayonix Face-ID SDK is based on Ayonix’s innovative facial recognition algorithms, which is especially designed for integration with any existing applications or completely new application. Ayonix algorithms are not optimized or trained on databases used for tests, like the FERET database. Ayonix uses its own internal proprietary databases, which do not contain data from test databases. Therefore, a realistic and stable face recognition is done.Verification Mode
1M:M (many-to-many) match
Identification mode is a comparison between many-entity and Many-entity. Perfect for recognition in a crowd in real-time.
21:M (one-to-many) match
Comparison between entities against to multiple entities. Perfect for scene where there is only one person in an entrance.
Technical Specifications
Item | Descriptions |
Input | Still Image File(BMP, JPG, PNG), IP Camera |
Speed | Face Detection: 100~200 milliseconds depending on scene complexity One-to-many matching: within half second for matching |
Template Size | Face template: 2Kbyte |
Database Size | No Limit |
Pose | Up to 30 degree. More than 30 degrees, there is a slight loss in matching ability. |
Race and Gender | Performs equally well on all races and both genders. |
Robustness to Variability | Robust with respect to changes in lighting conditions, expression, facial hair, hairstyle, Eyeglasses. |
Eyeglasses | Explicitly designed to match faces with or without eyeglasses, as long as the eyes are visible and not occluded by glare. |
Lighting | Optimal performance is obtained in diffuse ambient lighting, where the face is evenly illuminated, without shadows or glare. |
Background | Finds the faces in an image against any background, plain or cluttered but recommended to have a plain background. |
Image Depth and Resolution | Minimum of 320×240 resolution |
Head Size | Can recognize faces as small as 64x64 pixels or occupying 5% of the total image area. |