Why audio content analysis functionality has to be embedded?

It's been four years now

ORELIA is developing sound analysis systems for the detection of abnormal events, also referred as to Audio Content Analysis or Audio Analytics.

embedded audio analysis

To date, we have developed OSSR which performs noise sources (aircraft, cars, trains) recognition and helps better assess the impact of environmental noise. We also made a detection system of reverse noise allowing airports to identify the using of ultra-noisy aircraft braking system during landing. In the industry, ORELIA has developed an abnormal noise detector to identify defective products. Finally, in the field of security, ORELIA was one of the first company to introduce a fully configurable external module of sound analysis for immediate detection of events of interest : AUDIOSENSE 800.

The nature of sound is to warn a danger, it is omnidirectional, it avoids obstacles and is indifferent to lighting conditions. You hear a cry in the street and instantly you turn your head. The sound detection functionality as ORELIA conceived it through its technology can prevent an attack (verbal abuse, cry, assault, insult, call for help, rudeness), detect a degradation of material (broken glass) and alert from danger signal (fire alarm, barking dog ...).

In the field of security, audio analytics can be performed either on a server or on an external module placed near the audio sensor (similar to AUDIOSENSE 800) or directly into a camera or an intercom.

The first solution raises a huge problem named privacy : how to ensure that no conversation is recorded and replayed? The second solution solves the problem of privacy (only alarms are triggered outside of the module) but raises the problem of the cost generated by the installation of an additional module. This additional cost can be a real road block to the deployment of sound detection systems.

The third solution is to run the algorithms inside a camera or an intercom (audio analytics at-the-edge). It solves the two issues named respect for privacy and cost. Moreover, this configuration offers the advantage of alarms being triggered trough metatags inserted directly into the image. So there's no further integration in supervision platform needed.

Henceforth, ORELIA offers manufacturers advanced audio analytics functionalities. Now as an embedded software feature, Audio Analytics is cost effective and represents a competitive advantage.

For more information about  detection by means of edge audio analytics, write us at contact@orelia.fr


Additional information