Acoem to highlight new AI model for gunshot detection at GSX 2025

Acoem to highlight new AI model for gunshot detection at GSX 2025

Share this content

Facebook
Twitter
LinkedIn

Acoem has announced that it will showcase updates to its ATD gunshot detection technology and a new on-premises option for its Cadence software, at GSX.

According to the company, unlike legacy systems that depend on multiple sensors or remote servers, the Acoem ATD-300 Sensor System conducts detection and classification directly on the edge.

With ATD gunshot detection technology, pan-tilt-zoom (PTZ) cameras can immediately slew to the source of gunfire, giving first responders and security teams real-time situational awareness and captured evidence.

Acoem highlights that the ATD system delivers high accuracy with a single device, simplifying deployment and management across campuses, utilities and city environments.

Gunshot detection technology

Acoem says that the latest version of its AI model is built on a library of hundreds of thousands of qualified sounds collected over three decades, with up to 15,000 new sounds added monthly from live deployments.

Its engineering team, together with end-users through Cadence, continually validates and classifies noise events to refine detection.

This approach is designed to drive gunshot detection false positives toward zero while preserving unmatched accuracy in complex environments.

In addition to these regular updates, the ATD AI engine can reportedly be custom tuned for each deployment, learning and rejecting nuisance sounds unique to each individual environment.

“A recent example involved a gunshot sensor installed near a baseball field,” said Timothy English, Managing Director for Acoem ATD technology.

“The Acoem team was able to register the crack of the bat against the baseball with the neural net to ensure a home run is never interpreted as a gunshot, even though the sounds share very similar characteristics.”

Mission-critical deployments

At GSX, Acoem will also preview an on-premises version of its Cadence software. Until now, the Cadence software has been operated only as a cloud-based platform.

For mission-critical environments such as utilities, government and remote facilities, where outside internet connections are restricted for cybersecurity reasons, the new option will allow the software to run entirely inside an organization’s own network.

The Cadence software provides security teams with a single platform to easily manage ATD deployments.

It streamlines ATD-300 sensor setup and alarm notifications and verifies alerts with live audio clips for optional human validation through integrated video management software (VMS) systems.

“Security leaders are demanding both greater accuracy and greater deployment flexibility,” said English.

“With the new AI model, we’re raising the bar yet again on false positive rejection, and with on-premises Cadence software, we’re removing barriers for critical infrastructure sites that require closed-loop operations.”

Find out more at GSX at booth #3307.