MXSERVER’s flexible architecture supports multiple end-use applications which leverage a common core of cloud-based technology components coupled with “pluggable” machine learning services. The rapid growth, maturity and accuracy of artificial intelligence-based computer vision algorithms is generating new end-user applications that were previously unimaginable.
MXSERVER supports a myriad of intelligence and investigative use cases where large amounts of video and photo content has been collected and needs to be analyzed. Investigators batch load huge volumes of video surveillance and photograph evidence into the MXSERVER system, which automatically detects faces and indexes the content so it can be searched using facial recognition technology. MXSERVER uses advanced machine learning based face recognition technology to search all indexed videos and photographs to find additional instances where a suspect was seen. Investigators can then quickly piece together where and with whom potential suspects have been seen over time and at different locations. To determine the identity of a suspect, MXSERVER can also search digital “watchlist” databases containing pictures of previously known persons-of-interest.
MXSERVER and Matchbox deliver an enterprise grade physical access security capability which leverages machine vision and face recognition technology. Our advanced identity verification/face identification and video surveillance monitoring and exploitation technologies strengthen security and improve the protection of personnel at permanent installations and remote field locations (such as temporary security check points). MXSERVER and Matchbox deliver a “ubiquitous identification” capability – rapidly identifying an unknown person anywhere, anytime and using any video/digital camera with network connectivity.
Tygart’s Matchbox app enables remote users to quickly identify one or more individuals by simply using the embedded photo and video cameras on Android and iOS mobile devices (smart phones, tablets and Internet-of-Things devices). Using advanced machine learning algorithms, people’s faces are automatically detected, tracked and submitted to MXSERVER for identification in the cloud. Within seconds, match alerts are automatically displayed on the mobile device and in the cloud.
Tygart’s MXSERVER microservice computing framework enables organizations to rapidly operationalize machine learning algorithms at enterprise scale. New and innovative machine vision algorithms are being developed every day, but the benefits of such capabilities cannot be realized without coupling the algorithm with a high-performance processing platform capable of handling hundreds-of-millions of files. MXSERVER delivers a scalable and robust processing workflow management, distribution and storage capability which includes services such as: multiple batch file ingestion mechanisms, file management and storage, processing workflow management (job creation, prioritization, execution, error handling and process status tracking), data management and storage, alerting and notification mechanisms, and system health monitoring, administration and management.
At the center of Tygart’s MXSERVER™ system sits an algorithm and modality agnostic biometric search core which leverages Docker container-based microservices. The MXSERVER’s search core has been battle tested, delivering years of operational successes utilizing biometric galleries that scale to over a billion templates. Leveraging the latest machine learning based matching algorithms, our search system is able to achieve very high accuracy rates even when enrollment image quality is less than ideal. Tygart’s search core can be deployed in the cloud or on-premises and can be quickly integrated (via our robust web API) to biometric enable any application.
