devAIce®

Choose from our basic models or customized combinations. We can detect emotions from audio alone, or combined from audio and text. 

• Emotion classes (from audio): neutral, happy, angry, sad
• Emotion dimensions (from audio): activation, valence, dominance
• Multimodal emotion classes (from text and audio): neutral, happy, angry, sad
• Multimodal emotion dimensions (from text and audio): activation, valence, dominance

devAIce® can distinguish between background noises and human voice. It is very power efficient with low resource utilization when idle. The distinction between speech and music is available as an option.

• Age
• Gender

Distinction of up to 14 pre-defined acoustic scenes:

• Transportation: Car, Bus, Subway, Train
• Outside: Nature, Park, Street, Pedestrian Street, Large space
• Inside: Home, Bathroom, Public Building, Restaurant, Shop

Evaluate whether the speaker in a recording is the same as a previously-enrolled reference speaker.
Two modes are supported:

• Enrollment mode: a speaker model for a reference speaker is created or updated based on one or more reference recordings.
• Verification mode: a previously created speaker model is used to estimate how likely the same speaker is present in a given recording.
  
  

*Speaker Verification is currently in development and available in Beta version.

• Statistical descriptors based on pitch and loudness of voices.
• Speaking rate estimation in syllables per second.

Extract a selection of common openSMILE-based feature sets right from within devAIce.

devAIce® SDK:

• Platforms: Linux, Windows, macOS, Android, iOS, ROS
• Processor architectures: x86-64, ARMv7, ARMv8

devAIce® Web API:

• Accessible from any browser or cloud service via HTTP
• On-premise deployment as Linux virtual machine or Docker container

devAIce® is optimized for low resource consumption. Many models in devAIce® can run in real-time on embedded, low-power ARM devices such as Raspberry Pi and other SoCs.