SHORE® Software Solution – Technical Fact Sheet

Cross-Platform Software Library

  • Suitable for desktop, server and embedded systems
  • Runs on Windows, MacOS, Linux, iOS, Android
  • Supports x86, x64, ARMv7 and AARCH64 architectures
  • Further platforms are available on demand!

Development Platforms and Programming Languages

  • C++-SDK with an intuitive yet highly customizable API
  • Support for a wide range of development environments via bindings
    • Python (Download via Github)
    • Java (upon request)
    • Webassembly/Javascript (coming soon)
    • Support for other environments is available on demand!
    • No additional dependencies

Demo software

  • Free demo software available for Windows systems
  • Additional demo software is available upon request

Free demo software for Windows systems (additional demo software available on request!)

 

 

SHORE® – Classifications

Performance

SHORE® Benchmarks

The following benchmark results show the processing speed of SHORE® in frames per second on both desktop systems and embedded systems. The following parameters were varied:

  • Resolution of input images (VGA=640x480 pixels and HD=1280x720 pixels).
  • Number of faces per image (1 and 10)
  • 4 different configurations of SHORE®:
    • Face finding, determination of 68 facial landmarks and face analysis (6 facial expressions, age, gender)
    • Face finding and face analysis (6 facial landmarks, age, gender)
    • Face finding and determination of 68 Facial Landmarks
    • Face finding

 

A Raspberry Pi 3 Model B and an NVIDIA TX2 were used as reference systems for the measurements on embedded systems. The desktop system consisted of an Intel i7-7800k with 16 GB RAM.
The image data was processed in single frame mode - significantly higher frame rates can also be achieved when processing data with temporal correlation (e.g. camera recordings).

More Information

Face Detection: Fast, robust and in real-time with SHORE®

European Privacy Seal ePrivacyseal

The certification of the AVARD system has been in place since 2015.