The AMI Meeting Corpus: A Pre-announcement

Abstract

The AMI Meeting Corpus is a multi-modal data set consisting of 100 hours of meeting recordings. It is being created in the context of a project that is developing meeting browsing technology and will eventually be released publicly. Some of the meetings it contains are naturally occurring, and some are elicited, particularly using a scenario in which the participants play different roles in a design team, taking a design project from kick-off to completion over the course of a day. The corpus is being recorded using a wide range of devices including close-talking and far-field microphones, individual and room-view video cameras, projection, a whiteboard, and individual pens, all of which produce output signals that are synchronized with each other. It is also being hand-annotated for many different phenomena, including orthographic transcription, discourse properties such as named entities and dialogue acts, summaries, emotions, and some head and hand gestures. We describe the data set, including the rationale behind using elicited material, and explain how the material is being recorded, transcribed and annotated.

More Than Words: Inference of Socially Relevant Information from Nonverbal Vocal Cues in Speech

“…C2 contains all talk-shows (27 in total) broadcasted by Radio Suisse Romande during February 2005. C3 is the AMI meeting corpus [6], a collection of 138 meeting recordings involving 4 persons each and with an average length of 19 minutes and 50 seconds. The roles of C1 and C2 share the same names and correspond to similar functions: the Anchorman (AM), the Second Anchorman (SA), the Guest (GT), the Interview Participant (IP), the Headline Reader (HR), and the Weather Man (WM).…”

Detecting Action Items in Meetings

“…For these experiments, we use the AMI meetings corpus [1]. The corpus consists of about 100 hours of recorded and annotated meetings, divided into scenario and nonscenario meetings.…”

Audio Assisted Robust Visual Tracking With Adaptive Particle Filtering

“…Apart from "AV 16.3", we have also explored the suitability of several other publicly available audio-visual datasets, such as "CLEAR" [55], "AMI" [56] and "SPEVI" [57], and concluded that only the AV 16.3 dataset is suitable for the evaluation of our proposed methods. It complies with our requirements in terms of having circular microphone arrays with calibration information, mostly talking speakers, and challenging scenarios such as occlusion and rapid movements of the speakers.…”