LG Group subsidiary S&I Corporation this month has released its “atxpert (‘at’+’expert’) AI Viewer,” an AI building security management system that incorporates: Microsoft Azure; TnM Tech's Ghost-I AI edge computing platform; and Scenera's NICE PaaS, a managed Kubernetes service built on Microsoft Azure infrastructure.
S&I said that combining these partners’ technologies defines and enables a highly scalable and easy to manage building safety and security management system. According to the company, the system can detect special events such as off-hour intrusions, tailgating at the building's entrances, fire, loitering and more.
All such events are programmed to alert security managers with real-time, low latency notifications that will provide details of such events via apps running on mobile devices, prompting quicker reaction times. Cameras and X-ray scanners are connected to S&I's "atxpert AI Viewer" edge computing processor (Ghost-I) running edge AI analytics, enabling the motion detection with pre-programmed AI models.
S&I says the integrated system utilizes existing CCTV data of buildings, making it easy to be applied. As explained by the company, "the system recognizes, captures and filters only necessary events, as opposed to analyzing all streaming video images, allowing this model to be more cost-effective and saving on bandwidth costs, storage, and cloud-based video stream analytics."
Key features of the system, according to S&I, include:
- Built on existing safety management equipment – ONVIF compatible cameras and IoTs, and network infrastructure for low overhead system implementation.
- Scene-based event capture help organize events by easily searchable and indexable categories.
- On-premises edge computing to help reduce costs and latency of event detection and notification.
- Integration of X-Ray scanner with edge AI analytics for suspicious objects.
- Still image and short clips are immediately available upon event detection for viewing by security managers via apps running on mobile devices, prompting quicker responses.
- Accumulated and organized event data along with machine learning continues to improve AI models running on edge computing over time for higher accuracy.
- Deploying new applications and updating AI models are easily completed with managed Kubernetes services.
- High availability and low maintenance services.
- Security and privacy of data and analytics results.
- Offer other detection services for specific events such as tailgating at the entrance, personal incidents (falling), littering and others.
'atxpert AI Viewer' platform drill-down
S&I says it completed service development for atxpert AI Viewer in Q1 2021 and launched the service on buildings administered by LG Group, and is planning to further expand this service and apply it to more customers.
“Our expertise in building management is enhanced by the industry’s leading edge-to-cloud distributed AI services offered by our partners. Our commercialized services will elevate building management to the next level of security and safety,” said Mr. Eric Kim, Chief Director of S&I Corp. “The AI analytics data we gather in our service will further improve our capabilities, while also enabling us to integrate additional features. We plan to develop and offer safety services optimized on basis of AI-collected data of our sites."
According to S&I, Microsoft Azure is the foundation for the system and enables an edge to cloud pipeline managed by Scenera PaaS to process high bandwidth video streams effectively. Specifically, Azure Key Vault will be utilized to manage critical data security and privacy. Azure IoT Edge will be used to manage the deployment and reconfiguration of Edge AI Analytics. Azure IoT Hub will enable the management of thousands of edge devices.
The solution utilizes Azure Kubernetes Service managed by Scenera to handle large numbers of video and IoT data streams. AI algorithms trained using the Azure cloud infrastructure can be distributed from the cloud to the edge enabling the effective filtering of video data, while Azure Media Storages will store filtered and indexed data.
Bharat Sandhu, senior director, AI and Machine Learning at Microsoft Corp. commented, “At Microsoft we are committed to empower our partners and customers to innovate, and I am pleased to see this collaboration delivering a real value for S&I and its customers by enhancing building safety and security. This architecture can be applied to various applications beyond the building management solution including retail stores and large-scale residential property management.”
Scenera architected and built Platform as a Service (PaaS) in collaboration with Microsoft and has partnered with TnM Tech to provide end-to-end system integration and managed services for S&I.
David Lee, CEO of Scenera added, “We are glad to see our collaboration with Microsoft and TnM Tech produce an efficient and modern solution to S&I’s building management and look forward to seeing our services may implemented in other buildings and new businesses.”
Dr. ES Hwang, CEO of TnM Tech concluded, “We are glad to make it together with Microsoft and Scenera for S&I’s AI-based building security management service and will do our best to support it to be the most popular solution in the world.”