Deep Learning based Driver Monitoring Solutions
The immediate need towards Automotive Safety and In-Car Personalization has given rise to technology suppliers developing deep learning based driver monitoring solutions. AllGo has designed an array of solutions to help OEM’s and Tier 1’s build better and safer cars.

AllGo Owned Driver Monitoring IPs:


EYE GAZE TRACKING

  • Deep learning based iris and iris center tracking
  • Tracks eye state constantly and in real time
  • Using iris tracking, analyses gaze direction of the eyes
  • Support for eyes with all sizes, colors and shapes
  • Support camera at various angles
  • Support for GPUs

DRIVER DROWSINESS

  • The solution constantly monitors the eye status(open/close) and generates an alert when the drive is drowsy
  • Works with sunglasses, normal glasses, IR camera and normal camera
  • The solutions performance is at 8-10fps
  • Robust against illumination changes partial face occlusion, different kind of faces (gender, origin, skin colour)
  • Supported on ARM, DSP, GPU, ARM, x86
  • The algorithm’s efficiency is around 95%

FACE DETECTION

  • Deep learning based face detection
  • Detects and tracks faces in real time
  • Detects multiple faces in a single frame
  • Support for faces with different Ethnicities, Genders, Age, Expression, Partial occlusion and Illumination
  • Supports various angles of face (+/- 90o yaw, +/- 45o pitch, +/- 45o roll)
  • Support for GPUs

HEAD POSE ESTIMATION

  • Detects and tracks faces in real time
  • Detects multiple faces in a single frame
  • Support for faces with different Ethnicities, Genders, Age, Expression, Partial occlusion and Illumination

Also visit See ‘n Sense

Showcase

Drowsiness Detection
video1