Man-Machine Speech Communication

This book constitutes the refereed proceedings of the 17th National Conference on ManMachine Speech Communication, NCMMSC 2022, held in China, in December 2022.
The 21 full papers and 7 short papers included in this book were carefully reviewed and selected from 108 submissions. They were organized in topical sections as follows: MCPN: A Multiple Cross-Perception Network for Real-Time Emotion Recognition in Conversation.- Baby Cry Recognition Based on Acoustic Segment Model, MnTTS2 An Open-Source Multi-Speaker Mongolian Text-to-Speech Synthesis Dataset.

Klappentext

This book constitutes the refereed proceedings of the 17th National Conference on Man Machine Speech Communication, NCMMSC 2022, held in China, in December 2022. The 21 full papers and 7 short papers included in this book were carefully reviewed and selected from 108 submissions. They were organized in topical sections as follows: MCPN: A Multiple Cross-Perception Network for Real-Time Emotion Recognition in Conversation.- Baby Cry Recognition Based on Acoustic Segment Model, MnTTS2 An Open-Source Multi-Speaker Mongolian Text-to-Speech Synthesis Dataset.

Inhalt
MCPN: A Multiple Cross-Perception Network for Real-Time Emotion Recognition in Conversation.- Baby Cry Recognition Based on Acoustic Segment Model.- A Multi-feature Sets Fusion Strategy with Similar Samples Removal for Snore Sound Classification.- Multi-Hypergraph Neural Networks for Emotion Recognition in Multi-Party Conversations.- Using Emoji as an Emotion Modality in Text-Based Depression Detection.- Source-Filter-Based Generative Adversarial Neural Vocoder for High Fidelity Speech Synthesis.- Semantic enhancement framework for robust speech recognition.- Achieving Timestamp Prediction While Recognizing with Non-Autoregressive End-to-End ASR Model.- Predictive AutoEncoders are Context-Aware Unsupervised Anomalous Sound Detectors.- A pipelined framework with serialized output training for overlapping speech recognition.- Adversarial Training Based on Meta-Learning in Unseen Domains for Speaker Verification.- Multi-Speaker Multi-Style Speech Synthesis with Timbre and Style Disentanglement.- Multiple Confidence Gates for Joint Training of SE and ASR.- Detecting Escalation Level from Speech with Transfer Learning and Acoustic-Linguistic Information Fusion.- Pre-training Techniques For Improving Text-to-Speech Synthesis By Automatic Speech Recognition Based Data Enhancement.- A Time-Frequency Attention Mechanism with Subsidiary Information for Effective Speech Emotion Recognition.- Interplay between prosody and syntax-semantics: Evidence from the prosodic features of Mandarin tag questions.- Improving Fine-grained Emotion Control and Transfer with Gated Emotion Representations in Speech Synthesis.- Violence Detection through Fusing Visual Information to Auditory Scene.- Mongolian Text-to-Speech Challenge under Low-Resource Scenario for NCMMSC2022.- VC-AUG Voice Conversion based Data Augmentation for Text-Dependent Speaker Verication.- Transformer-based potential emotional relation mining network for emotion recognition in conversation.- FastFoley Non-Autoregressive Foley Sound Generation Based On Visual Semantics.- Structured Hierarchical Dialogue Policy with Graph Neural Networks.- Deep Reinforcement Learning for On-line Dialogue State Tracking.- Dual Learning for Dialogue State Tracking.- Automatic Stress Annotation and Prediction For Expressive Mandarin TTS.- MnTTS2 An Open-Source Multi-Speaker Mongolian Text-to-Speech Synthesis Dataset.