AMI Meeting Corpus

The AMI Meeting Corpus is a multi-modal data set comprising 100 hours of meeting recordings. It has been meticulously curated for research purposes and includes various modes of data capture. Let me provide you with more details:

1. Purpose and Context:

   • The corpus was created in the context of a project that aims to develop meeting browsing technology.
   • Eventually, it will be publicly released for use by researchers and practitioners.

2. Data Composition:

   • 100 hours of recorded meetings are included.
   • The data is collected from various domains and scenarios.
   • Around two-thirds of the data involves participants playing different roles in a design team, taking a design project from kick-off to completion over the course of a day.
   • The remaining data consists of naturally occurring meetings.

3. Modalities and Annotations:

   • The corpus includes synchronized recording devices such as close-talking and far-field microphones, individual and room-view video cameras, projection, a whiteboard, and individual pens.
   • Annotations cover various phenomena, including orthographic transcription, dialog acts, and head movement.