There is a newer version of the record available.

Published December 14, 2023 | Version v1
Dataset Restricted

IEEE ICME 2024 Grand Challenge: Semi-supervised Acoustic Scene Classification under Domain Shift

Authors/Creators

  • 1. ROR icon Northwestern Polytechnical University
  • 2. ROR icon Institute of Acoustics
  • 3. ROR icon Chinese Academy of Sciences
  • 4. ROR icon University of Surrey
  • 5. ROR icon Nanyang Technological University

Description

Chinese Acoustic Scenes 2023 is a large-scale dataset that serves as a foundation for research related to environmental acoustic scenes. The dataset includes 10 common acoustic scenes, with a total duration of over 7800 minutes (approximately 130 hours). Each audio clip is 10 seconds long and provides metadata about the recording location and timestamp. The dataset was collected by members of the Joint Laboratory of Environmental Sound Sensing (JLESS) at the School of Marine Science and Technology, Northwestern Polytechnical University. The data collection spanned from April 2023 to September 2023, covering 22 different cities across China.

For the ICME 2024 Grand Challenge - Semi-supervised Acoustic Scene Classification under Domain Shift, we will provide a development dataset with around 24 hours based on the CAS 2023 dataset. Specifically, the dataset will contain a small amount of labeled data (around 4 hours) and a large amount of unlabeled data (around 20 hours). Approximately 20% of the labeled data will be allocated for validation. Moreover, around 3 hours of unseen data will be provided for evaluation.

Files

Restricted

The record is publicly accessible, but files are restricted to users with access.