Machine Learning for Multimodal Interaction: First International Workshop, MLMI 2004, Martigny, Switzerland, June 21-23, 2004, Revised Selected Papers ... Applications, incl. Internet Web, and HCI)

This book constitutes the thoroughly refereed post-proceedings of the First International Workshop on Machine Learning for Multimodal Interaction, MLMI 2004, held in Martigny, Switzerland in June 2004. The 30 revised full papers presented were carefully selected during two rounds of reviewing and revision. The papers are organized in topical sections on HCI and applications, structuring and interaction, multimodal processing, speech processing, dialogue management, and vision and emotion. Table of Contents......Page 10 Accessing Multimodal Meeting Data: Systems, Problems and Possibilities......Page 14 Browsing Recorded Meetings with Ferret......Page 25 Meeting Modelling in the Context of Multimodal Research......Page 35 Artificial Companions......Page 49 Zakim – A Multimodal Software System for Large-Scale Teleconferencing......Page 59 Towards Computer Understanding of Human Interactions......Page 69 Multistream Dynamic Bayesian Network for Meeting Segmentation......Page 89 Using Static Documents as Structured and Thematic Interfaces to Multimedia Meeting Archives......Page 100 An Integrated Framework for the Management of Video Collection......Page 114 The NITE XML Toolkit Meets the ICSI Meeting Corpus: Import, Annotation, and Browsing......Page 124 S-SEER: Selective Perception in a Multimodal Office Activity Recognition System......Page 135 Mapping from Speech to Images Using Continuous State Space Models......Page 149 An Online Algorithm for Hierarchical Phoneme Classification......Page 159 Towards Predicting Optimal Fusion Candidates: A Case Study on Biometric Authentication Tasks......Page 172 Mixture of SVMs for Face Class Modeling......Page 186 AV16.3: An Audio-Visual Corpus for Speaker Localization and Tracking......Page 195 The 2004 ICSI-SRI-UW Meeting Recognition System......Page 209 On the Adequacy of Baseform Pronunciations and Pronunciation Variants......Page 222 Tandem Connectionist Feature Extraction for Conversational Speed Recognition......Page 236 Long-Term Temporal Features for Conversational Speech Recognition......Page 245 Speaker Indexing in Audio Archives Using Gaussian Mixture Scoring Simulation......Page 256 Speech Transcription and Spoken Document Retrieval in Finnish......Page 266 A Mixed-Lingual Phonological Component Which Drives the Statistical Prosody Control of a Polyglot TTS Synthesis System......Page 276 Shallow Dialogue Processing Using Machine Learning Algorithms (or Not)......Page 290 ARCHIVUS: A System for Accessing the Content of Recorded Multimodal Meetings......Page 304 Piecing Together the Emotion Jigsaw......Page 318 Emotion Analysis in Man-Machine Interaction Systems......Page 331 A Hierarchical System for Recognition, Tracking and Pose Estimation......Page 342 Automatic Pedestrian Tracking Using Discrete Choice Models and Image Correlation Techniques......Page 354 A Shape Based, Viewpoint Invariant Local Descriptor......Page 362 S......Page 374 Z......Page 375

This book constitutes the thoroughly refereed post-proceedings of the First International Workshop on Machine Learning for Multimodal Interaction, MLMI 2004, held in Martigny, Switzerland in June 2004.

The 30 revised full papers presented were carefully selected during two rounds of reviewing and revision. The papers are organized in topical sections on HCI and applications, structuring and interaction, multimodal processing, speech processing, dialogue management, and vision and emotion.

Several large-scale projects (e.g. [1,2]) have examined the collection, analysis and browsing of multimodal meeting data.