Skip to main content

3D Face Mesh Tracking

Edge Enabled Facial Tracking for Realistic Digital Experiences.

Unlock the power of our advanced 3D Face Mesh Tracker, designed to bring digital avatars, augmented reality (AR), health insights, privacy-aware solutions, retail engagement, and real-time expression analysis to life.

Explore the Possibilities

Designed for real-time performance, it seamlessly translates facial movements into dynamic, high-fidelity three-dimensional digital representations.

Key Features

High Accuracy

Captures detailed facial geometry and expressions with precision.

Life-like Realism

Generates natural and expressive animations for digital avatars.

Optimized for Cost-Effective System on Chips (SoCs)

Runs efficiently on resource-constrained devices without compromising performance.

Multi-Platform Compatibility

Supports integration across various platforms and operating systems.

Lightweight 2.5 MB Model Size

Compact yet powerful model ensures quick deployment and minimal storage requirements.

Edge Optimized Performance

Runs in real time on-device for low latency, improved privacy, and reliable offline operation.

High-Speed Processing

Real-time face tracking with up to 68 frames per second (FPS) on CPU and 160 FPS on GPU for seamless user experiences.

Advanced 3D Mesh for Realism and Precision

AlgoFace’s 3D Facial Mesh Tracker uses an advanced blendshape method to bring avatars to life. With 1119 vertices, 2128 triangles, and detailed semantic landmarks, the SDK provides an intricately detailed, responsive 3D mesh. Realistically reflecting facial traits and expressions, this technology transforms face motion into seamless animation.

Advanced 3D Face Tracking

Our 3D FLT Mesh captures a broad range of head poses and facial movements:

  • Dense Facial Landmarks: Tracks 1119 3D points for high-fidelity capture of facial expressions and head movement.
  • Head Pose: Wide range of yaw, pitch, and roll.
  • Axis Translation: Real-time tracking along x, y, and z axes.
  • Customizable Camera Parameters: Includes focal length and camera center adjustments to adapt to various setups.
  • Additional Semantic Facial Landmarks: For mapping common facial features like the corners of the eyes, pupils, edges of the lips, tip of the nose, and contour of the jawline.

Seamless Data Integration

Image Input Formats:

  • bmp
  • jpeg, jpg, jpe, jp2
  • png
  • pbm, pgm, ppm
  • tiff, tif

Video Input Formats:

  • mp4
  • avi
  • mov
  • divx
  • mpeg

Output Data:

Raw numpy format with detailed facial mesh and semantic landmarks.