Multi-Speaker
Machine Listening Skill
Recognize multiple speakers, identify languages, emotions & intents from natural conversations (meetings, interviews, calls).
Enrich your ASR with DeepAffects
1. extract HUMAN CONVERSATIONS FROM AUDIO
2. rECOGNISE SPEAKERS, identify EMOTIONS & INTENTS
3. combine deepaffects insights WITH ASR to enhance the ENTERPRISE APPLICATIONS
Multi-Speaker Machine Listening for Voice-Apps
CALL CENTER APPS
Identify agent & customer voice, sentiments & intents to measure customer satisfaction & generate personalized agent coaching.
HEALTHCARE APPS
Extract prosodic features to build high specificity & sensitivity deep neural nets based models to screen & monitor patients for mental & respiratory illness.
TRANSCRIPTION APPS
Enhance accuracy of the interview, conference, meeting transcripts by removing noise and recognizing multiple speakers.
MEETING APPS
Recognize multiple speakers & their intents in a meeting to monitor dynamics and equity of contribution/participation in meetings.
Our approach differentiates us
Speech Enhancement
Spectral Analysis VS Speech-to-text
Multi-Speaker (>= 2) in the wild
multi-lingual
Short & Long utterance
Robust DNN based Features
Large Training dataset
Balanced training dataset to remove bias
Clustering algorithm with Anomaly Detection
Domain specific Intent Detection
infinitely scalable
secured
