Eluviant launches video AI model for enterprise surveillance
New technology analyzes motion sequences in real time to identify workplace safety hazards and security incidents.
Artificial intelligence developer Eluviant, formerly known as IntelexVision, has introduced Aurora Flow, a new video understanding model designed for large-scale enterprise surveillance networks.
The system is designed to address a common bottleneck in industrial and commercial security operations: the difficulty of monitoring hundreds of live camera feeds simultaneously. By analyzing continuous sequences of movements rather than individual video frames, this technology identifies complex behavioral patterns as they occur.
Traditional computer vision tools typically flag objects or static scenes, which can lead to a lack of context. The new model processes chronological sequences of movement over time and allows the system to automatically flag certain physical behaviors, such as climbing, physical altercations, and theft, that previously required continuous human observation to detect.
Built for high-security environments, this software can operate entirely on-premises and within air-gapped networks. That is, it works without an active connection to the external Internet.
The software builds on the company’s existing technology suite, including an unsupervised self-learning system and vision language models that have been utilized in live monitoring operations for over a year.
According to company data, the underlying platform is currently in use in more than 250 active deployments, monitoring approximately 50,000 camera feeds worldwide. Common applications for this technology span critical infrastructure, transportation hubs, smart cities, and heavy industrial facilities where real-time incident response is required.
Along with the product release, the company completed a corporate rebranding process, transitioning from its previous name IntelexVision to Eluviant.
About the author
Jesse Jacobs is an associate editor at OHSOnline.com.
