Project Metior Telum (Latin for ‘measure the weapon’) is building a large-scale automatic firearm detection system by creating a hybrid machine learning dataset – combining real images with 3D produced images of firearms (scanned from real AFP firearms).

The dataset is being developed by rendering the scanned firearms in various lighting, angles, conditions, and then compositing them into different backgrounds. This makes sure that the synthetic images produced reflect how real-world firearms appear in actual operational settings.

Through this, we are able to ethically and rapidly source data to develop AI solutions with accurate firearm detection.

Metior Telum will also be used to advance Nightwatch.