Face Detection Software SHORE®: Fast, Reliable and Real-time Capable

Pioneering Image Analysis

The software library SHORE® allows for the quick detection of faces in image sequences and videos. It can estimate gender, age and facial expressions.

Your benefits at a glance:

Face detection and Analysis

  • Real time detection of faces
  • Real time detection of facial features like eyes and mouth and their state (open/closed)
  • Estimation of gender and age
  • Estimates four facial expressions: happy, sad, surprised, angry

Technical Key Features

  • Works offline: no internet connection required
  • Runs on standard platforms as well as mobile devices
  • Very low hardware requirements
  • Privacy by Design

Various Application Areas

  • Market research
  • Cognitive robotics
  • Health Support
  • Automotive
  • And many more

An extension of SHORE® consisting of a camera-based pulse recognition is currently being tested.

Conform to German General Data Protection Regulation – Interview with Jens Garbas

Jens Garbas is an expert in the image analysis software SHORE®. Data protection is crucial when using software-based facial analysis. In this interview, he provides an insight into how SHORE® works, its potential applications, and the measures taken to protect data.

Privacy warning

With the click on the play button an external video from wwww.youtube.com is loaded and started. Your data is possible transferred and stored to third party. Do not start the video if you disagree.

SHORE® Benchmarks

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


• Resolution of the input images (VGA = 640 × 480 pixels and HD = 1,280 × 720 pixels)
• Number of faces per image (1 and 10)
4 different configurations of SHORE®:

            Face detection, determination of 68 facial landmarks and face analysis (6 facial expressions, age, gender)
            Face detection and face analysis (6 facial expressions, age, gender)
            Face detection and determination of 68 facial landmarks
            Face detection     


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.
Image data were processed in single-image mode – and significantly higher frame rates were achieved when processing data with a time correlation (e.g. images recorded with a camera).

 

        

 

      

Privacy by Design

The report published on June 8th 2017 by the Bavarian Data Protection Authority for the Private Sector (BayLDA) confirms that the AVARD system doesn’t process personal data but only anonymous data. Thus, AVARD ensures the compliance with data protection laws at 100%.

 

For more data protection

In October 2017 the AVARD system has successfully passed a recertification audit to verify compliance with German or European data protection legislation.

FAQ

 

Data protection with the SHORE® image analysis software

The software complies with the German General Data Protection Regulation – Interview with Jens Garbas

Video

SHORE® in action

 

Brochure

SHORE® – Pioneering Facial Analysis

 

Press Releases and Web News

 

Take a look at the various application fields of SHORE®

SHORE® on Google Glass