FaceMatchR gives you the ability to access our facial biometrics engine on the go. You can match images from your photo gallery, Facebook and Instagram account. This is a demonstration app to bring facial recognition to the public on a mobile platform.
This version only allows you to perform image matching with a 8 face limit for each picture. This will give you a maximum of 8x8 facial recognition comparisons per pair of images.
HOW IT WORKS
Once the user has selected 2 images with faces, faces are detected and cropped on the mobile device. These cropped images are then sent to our back end cluster through our customized API and matched 1:1 against each other. The cluster queues up and processes the faces in batch jobs. The job queue is distributed between 2 commodity servers that execute the facial recognition matching algorithm. The cluster gives the results back to the device through MQTT messages.
FEATURES AT A GLANCE
- Facial detection of images
- Facial recognition of faces in images
- Group picture face detection and recognition
- Source images from social media accounts
- Fast real time image processing
- Backend cluster image processing
- low data bandwidth requirements
DETAILS
FaceMatchR allows you to test our facial recognition cluster.
1. You select 2 images from one of the three available sources (photo gallery, Facebook, and Instagram).
2. Facial detection is performed on device.
3. Cropped faces are uploaded to our cluster.
4. Facial recognition algorithm performed on cropped faces (1-N).
5. Results are sent back to device.
ABOUT US
We are a Research and Development company currently focused on biometrics. In 2014 - 2015 we implemented a nation wide customized security solution for the First European Olympic Games, Baku 2015. The system included access control, asset tracking, surveillance and facial recognition. Visit our website Prilyx.com for more details about us.
Feel free to contact us for more information on facial recognition or development of a customized system.