Sound Intelligence' processing techniques are based on a linear model of the human inner ear (cochlea). The cochlea model performs a frequency analysis of the incoming sound. The advantage of the cochlea model is that it guarantees the continuity in both time and the frequency domain.
Sound Intelligence' analysis techniques are derived from a knowledge based approach. This means that temporal and spectral information is used for the analysis of audio. Advantages are that no large amounts of training samples are needed and that the system has no problems with low signal-to-noise ratios. Therefore the SI approach is very suitable for common unconstrained environments like prisons, city centers and public transport.
Sound Intelligence' technology can easily be integrated with other systems, like camera management systems, video analytics systems, intercoms, condition monitoring systems, callcenter applications, etc. An open programmable message-based interface is available.
Central or 'on the edge'?
The SI technology can run on a central server for analysis of multiple audio channels, or on a one-channel 'edge device'. Both analog and IP audio inputs are supported.