FacialSensing

The facial recognition software FaceU can detect the position of faces in an image, as well as the feature points of each face (such as contours, eyes, nose, mouth, etc.). It can also estimate face orientation and the direction of the gaze. Additionally, it has a facial authentication function to estimate the identity of individuals, as well as functions to estimate the apparent age and gender.

Examples

Examples

Driver Monitoring

  • Detects driver distraction and drowsiness
  • Identifies the driver and automatically adjusts the seat position

Store Visitor Analysis

  • Counts the number of visitors and estimates their age and gender distribution

Digital Signage Viewing Analysis

  • Counts the number of people who viewed the digital signage and estimates their age and gender

Facial Authentication for Employee Management

  • Restricts PC usage to authorized personnel only

Game Character Transformation

  • Transforms faces into similar characters based on facial feature points
Main Function

Main Function

  • Face Detection: Four-point face bounding box, up to 20 faces simultaneously
  • Facial Feature Point Detection: Up to 72 points
  • Face Orientation Detection: Yaw, Pitch, Roll
  • Gaze Detection: Yaw, Pitch
  • Facial Authentication: Feature extraction, feature matching
  • Spoofing Detection: Determines if the input face image is a spoof
  • Age Estimation: Estimates age from 3 to 99 years in one-year increments
  • Gender Estimation: Outputs male or female
  • Others: Head position estimation, PERCLOS
Features

Features

  • Face Detection, Facial Authentication, and Age/Gender Estimation are possible even for faces wearing masks.
  • Considering the use in embedded SoCs, we ensure good recognition accuracy even under certain hardware resource constraints.
  • Enhanced environmental robustness against changes in light and background, allowing accurate face recognition in images captured in highly variable lighting conditions within moving objects.
Offerings

Offerings

  • We offer two lineups according to usage:
    General Use:
    Suitable for scenes with one to several dozen faces.
    Detailed Monitoring of Specific Faces:
    Ideal for scenes requiring detailed monitoring of a few faces, including support for near-infrared images.
  • We provide the recognition library (SDK), sample programs for operation verification, and technical documents such as API specifications.
  • Supported platforms by default are Windows, PC-Linux, and ARM-Linux, but we can also accommodate other platforms based on specific needs.
product