Sample dataThis is a short sample of 5 minutes of an "Interactive Seminar" from one lab. It contains the data, the video labels and the transcriptions as it will be present in the development set. The setup consists of: - 5 cameras ( 4 angle cameras and one eye-fish camera)
- 1 Mark III microphone array
- 3 to 7 T-shaped arrays
- room information like background pictures, calibrations pictures and values, schematics,...
The test set will contain only the data themselves. The labels are in the archive: labels.tar.gz [10M]
The video data are in the archive: video.tar.gz [1.5G] The audio data from the T-shaped arrays are in the archive: audio_Tarray.tar.gz [815M]
The audio data from the Mark III array are in the archive: audio_MarkIIIarray.tar.gz.part1.tar [1.2G] and audio_MarkIIIarray.tar.gz.part2.tar [961M] To extract this last splitted archive: C:\tar>tar -x -M --file=audio_MarkIIIarray.tar.gz.part1.tar
Prepare volume #2 for `audio_MarkIIIarray.tar.gz.part1.tar' and hit return: n audio_MarkIIIarray.tar.gz.part2.tar Prepare volume #2 for `audio_MarkIIIarray.tar.gz.part2.tar' and hit return:
C:\tar>tar -xzf audio_MarkIIIarray.tar.gz
Further Documentation For a complete description of the CHIL - CLEAR Data and annotations see the following document: The CHIL Audiovisual Corpus for Lecture and Meeting Analysis inside Smart Rooms.
| 
