NorthStar Scenario - Autonomous Vision Intelligence
At his workplace, Jonathan, head of the workplace “safety and management” of a retail franchise, is equipped with an omnipotent surveillance copilot.
“Starting this noon the hallway in building A will be under construction. can you start monitoring the zone and send alert level 3B when someone enters it without wearing a helmet and mask?”
Not only will Jonathan receive email alerts, but he will also automatically get a link to the dashboard for detailed summaries. This retail brand owns more than 500 stores nationwide, most of which have video cameras deployed at key locations such as entrances and parking lots. With this new copilot, Jonathan also helps other departments discover business insights. For instance:
By indicating "Starting from now, record the peak times and the utilization of parking lots." Or even "Collect basic data in domain of a retail shop monitoring" and letting the system translate "peak time" or "utilization" to the basic data, ModelPlus will collect so that later anyone authorized can ask: “tell me how the peak times in midwest top cities have changed in last season” and “summarize the utilization of parking lots for our busiest and most idle stores”. In this way Jonathan’s system will save the analytic team many days of time.
Zooming out from this specific story is the technological revolution it represents. The picture above shows the evolving path we see for vision intelligence systems. A decade ago, the concept of Model-as-a-Service and its incarnations brought intelligent vision systems to a new generation by decoupling MLOps (Machine Learning Operations) from business logic. Now we face another opportunity to remove even more manual effort for adapting and integrating various models, to break the data silo with multi-modal large models, and provide superior human interactions. We envision this wave of innovations to prepare for the true autonomous Robotic era by building a solid foundation for cognition and decision making.
Back to homepage
Back to News page