Chinese bulbul

Kramxel sound classifier is a low memory footprint, fast learning AI engine, designed for offline applications. It can learn sounds of different bird species after only dozens of training samples. The learning engine can produce very low footprint classifier works even on IoT SoC. The application of sound classifier is unlimited:

For home security, it can be used to identify dog bark, crying and alarm bells, to send alert or trigger recording events.

For wild camera, it can be used to start recording only when specific animal is in front of camera.

It is also possible for user to create their own class of sound, and provide just a few samples, the engine can learn it and use it right away.

classifier training

Kramxel provide test account for potential customers to test sound classifier, the test site use exactly the same algorithm as in embedded system, it will show identical result as the sound is received in the device.

Please contact us if you have application for it and need a test account.