The Fraunhofer Institute for Integrated Circuits IIS has adapted its SHORE real-time face detection and analysis software for the Google Glass. The app detects people’s faces and determines their emotions by analyzing facial expressions.
SHORE is a software platform that is the result of years of research and development. The software runs in real time and is able to detect faces down to a minimum size of 8 x 8 pixels. When the Google Glass’s 5 MPixel camera captures an image, it is paired against a database of 10,000 annotated faces and processed using structure-based features and learning algorithms. With SHORE, Fraunhofer researchers were able to train so-called models that boast extremely high recognition rates. Specifically, it features a 91.5% front facial recognition detection rate, a 94.3% gender detection rate, and a 6.85 annual mean absolute error rate for age estimation.
All calculations are performed in real-time by the CPU integrated in the eyewear and the image data is stored in the device. SHORE for Google Glass is the first emotion recognition software to function in real-time with the wearable device. This technology opens up a new spectrum of smart eyewear applications, including communication aids for people with disorders such as autism, many of whom have difficulty interpreting emotions through facial expressions. Any missing information could be superimposed in the wearer’s field of vision with data glasses. Even the visually impaired can benefit from the new software by receiving supplementary audio information about people in their surroundings, according to Fraunhofer.
Fraunhofer was able to develop this app due to its participation in the Google Glass “Explorer Program,” during which the SHORE software library was adapted and implemented as Glassware.
View the Fraunhofer press release.
Share your vision-related news by contacting James Carroll, Senior Web Editor, Vision Systems Design
To receive news like this in your inbox, click here.