MXSERVER™ and Matchbox® leverage cloud-technology and deep learning-based face detection and recognition algorithms to:
- Rapidly analyze and search massive collections (hundreds of millions) of videos/photos, and
- Monitor large arrays of video surveillance cameras in real time
Combined, MXSERVER and Matchbox deliver “ubiquitous identification” – the ability to rapidly identify an unknown person anywhere, anytime and using any face capture device and/or video/digital camera which has Internet or LAN connectivity. MXSERVER and Matchbox are enabling soldiers, security personnel, forensic examiners, intelligence analysts, and law enforcement investigators to:
- Identify persons-of-interest using face recognition
- Continuously monitor large arrays of video surveillance devices to detect and identify persons of interest
- Identify, locate, and track persons of interest, their associates and their activities across time and locations
- Identify individuals in field environments using smart phones – including the ability to quickly scan crowds for threats using a mobile device’s embedded video camera
- Manage cameras and other sensors, watchlists, and face indexed video/photo repositories in the cloud
- Generate, manage and transmit match notifications and alerts in the cloud
- Collect and search Social Media content using face detection and face recognition
- Rapidly add new video analytic algorithms into MXSERVER’s highly flexible and scalable video processing workflow framework
MXSERVER’s flexible architecture supports multiple end-use applications which leverage a common core of cloud-based technology components. The platform has been proved to scale to unprecedented levels in the industry – indexing and searching hundreds-of-millions of videos/photos containing billions of face images.

Tygart’s MXSERVER platform delivers a highly flexible and scalable microservice oriented high-performance computing platform into which a broad spectrum of machine learning algorithms (biometric detection and matching, object detection and classification, image enhancement and manipulation, etc.) can be rapidly integrated.