This book constitutes the thoroughly refereed post-proceedings of the Third International Workshop on Machine Learning for Multimodal Interaction, MLMI 2006, held in Bethesda, MD, USA, in May 2006. The papers are organized in topical sections on multimodal processing, image and video processing, HCI and applications, discourse and dialogue, speech and audio processing, and NIST meeting recognition evaluation.

MLMI'06.- Model-Based, Multimodal Interaction in Document Browsing.- The NIST Meeting Room Corpus 2 Phase 1.- Audio-Visual Processing in Meetings: Seven Questions and Current AMI Answers.- A Multimodal Analysis of Floor Control in Meetings.- Combining User Modeling and Machine Learning to Predict Users' Multimodal Integration Patterns.- Using Audio, Visual, and Lexical Features in a Multi-modal Virtual Meeting Director.- A Study on Visual Focus of Attention Recognition from Head Pose in a Meeting Room.- Multi-person Tracking in Meetings: A Comparative Study.- Gaussian Mixture Models for CHASM Signature Verification.- Kalman Tracking with Target Feedback on Adaptive Background Learning.- Da Vinci's Mona Lisa.- The Connector Service-Predicting Availability in Mobile Contexts.- Multimodal Input for Meeting Browsing and Retrieval Interfaces: Preliminary Findings.- Gesture Features for Coreference Resolution.- Syntactic Chunking Across Different Corpora.- Multistream Recognition of Dialogue Acts in Meetings.- Text Based Dialog Act Classification for Multiparty Meetings.- Detecting Action Items in Multi-party Meetings: Annotation and Initial Experiments.- Overlap in Meetings: ASR Effects and Analysis by Dialog Factors, Speakers, and Collection Site.- A Speaker Localization System for Lecture Room Environment.- Robust Speech Activity Detection in Interactive Smart-Room Environments.- Automatic Cluster Complexity and Quantity Selection: Towards Robust Speaker Diarization.- Speaker Diarization for Multi-microphone Meetings Using Only Between-Channel Differences.- Warped and Warped-Twice MVDR Spectral Estimation With and Without Filterbanks.- Robust Heteroscedastic Linear Discriminant Analysis and LCRC Posterior Features in Meeting Data Recognition.- Juicer: A WeightedFinite-State Transducer Speech Decoder.- Speech-to-Speech Translation Services for the Olympic Games 2008.- The Rich Transcription 2006 Spring Meeting Recognition Evaluation.- The IBM RT06s Evaluation System for Speech Activity Detection in CHIL Seminars.- A Lightweight Speech Detection System for Perceptive Environments.- Robust Speaker Diarization for Meetings: ICSI RT06S Meetings Evaluation System.- Technical Improvements of the E-HMM Based Speaker Diarization System for Meeting Records.- The AMI Speaker Diarization System for NIST RT06s Meeting Data.- The 2006 Athens Information Technology Speech Activity Detection and Speaker Diarization Systems.- Speaker Diarization: From Broadcast News to Lectures.- The ISL RT-06S Speech-to-Text System.- The AMI Meeting Transcription System: Progress and Performance.- The IBM Rich Transcription Spring 2006 Speech-to-Text System for Lecture Meetings.- The ICSI-SRI Spring 2006 Meeting Recognition System.- The LIMSI RT06s Lecture Transcription System.