Biometric facial Image autocapture, quality assurance, and camera abstraction

PreFace is an SDK that automatically captures and analyzes biometric facial images in order to maximize their quality and matchability. It can enable a biometric enrollment application to automate the facial image capture process and also ensure that enrolled images comply with ISO standards or backend processing system and are of sufficient quality to perform biometric matching.

PreFace integrates with the camera to perform analysis of the live facial image stream or video. Once basic quality criteria are met, PreFace triggers the camera to take a full-resolution image. Following capture, PreFace performs a thorough image analysis, which reports image geometry, non-compliant features, and spoof attempts. Scaling, rotation, and cropping of the image is performed to meet highly configurable targets and thresholds. These thresholds are derived from the ISO/IEC 19794-5 standard for biometric facial image quality. Results are reported to the user. PreFace also includes a robust face finder, able to locate multiple faces in a single frame in both still shots and video.

PreFace Screenshot

PreFace Screenshot


Features and Functionality

Camera API

PreFace includes “Camera API,” which serves to abstract camera hardware and integrate software-driven autocapture with a variety of consumer-grade digital cameras, webcams, and industrial cameras. It is designed to greatly simplify the task of integrating a facial image camera into a photo capture application.  Camera API provides a method by which to support many different cameras within a single application: program once, and use with many camera models. Support for new cameras is added in subsequent revisions of the SDK as they become available. Camera API enables a biometric application to operate equivalently with a variety of devices over time or within the same system. Cameras are also tested and submitted for approval by the GSA for use in FIPS 201 compliant “PIV” U.S. government employee credentialing systems. An up-to-date list of cameras supported by Camera API is available from Aware upon request.

PreFace functionality includes:

  • Pose Angle Yaw
  • Dynamic Range
  • Brightness
  • Saturation
  • Smile
  • Eye Contrast
  • Eye Obstructed (Left or Right)
    – Glasses Frames
    – Hair
    – Closed Eye
    – Eye Valid
  • Off-angle Gaze
  • Red-eye
  • Number of Image Channels
  • Background
    – Gray
    – Uniformity
    – Clutter
    – Type
    – Color Balance
    – Pad Type
  • Conditional Padding
  • Illumination Asymmetry
  • Facial Shadows
  • Focus
  • Sharpness
  • Unnatural Skin Color
  • Glasses
  • Glasses with Dark Lenses
  • Glasses Glare
  • Glasses with Heavy Frames
  • Forehead Obstructed
  • Image Width
  • Image Height
  • Eye Separation
  • Eye Axis Location Ratio
  • Centerline Location Ratio
  • Image Height/Width Ratio
  • Image Width/Head Width Ratio
  • Head Height/Image Height Ratio
  • Eye Axis Angle
  • Estimated Age
  • Gender-Female
  • Gender-Male
  • Race-White
  • Race-Black
  • Race-Asian
  • JPEG Quality Level
  • File Size
  • JPEG2000 Compression Ratio
    – Within ROI
    – Outside ROI
  • Image Format