Speech Processing in Everyday Environments (CHiME 2018)
7th September, Microsoft, Hyderabad (a satellite event of Interspeech 2018)
The workshop took place in September 2018. This website is a record of the event providing access to the papers, posters and presentations. The proceedings will also be available via the ISCA online archive.
- training schemes: data augmentation, semi-supervised training,
- speaker localization and beamforming,
- single- or multi-microphone enhancement and separation,
- robust features and feature transforms,
- robust acoustic and language modeling,
- robust speech recognition,
- robust speaker and language recognition,
- robust paralinguistics,
- cross-environment or cross-dataset performance analysis,
- environmental background noise modelling.
Papers reporting evaluation results on the CHiME-5 dataset or on other datasets are equally welcome.
The CHiME-5 ChallengeAs a focus for discussion, the workshop will host the 5th CHiME Speech Separation and Recognition Challenge (CHiME-5). The new challenge will consider the problem of distant microphone conversational speech recognition in everyday home environments. The challenge will target multi-microphone solutions with separate tracks for fixed microphone arrays and head-mounted microphones. Speech material will be elicited using a dinner party scenario with efforts taken to capture data that is representative of natural conversational speech. To find out more please visit Challenge Website.
CHiME 2018, SeptemberHyderabad, India
- Jon Barker &
Univ. of Sheffield, UK
- Shinji Watanabe Johns Hopkins Univ., USA
- Emmanuel Vincent
- More info
- Simerpreet Kaur Microsoft
- Registration: Jun 8th
- Submission: Aug 3rd
- Notification: Aug 20th
- Workshop: Sep 7th
- More info