Audio Signal Dataset for Danger Detection of Women and Children
Description
This publicly available audio signal dataset is built to detect danger for women and children. It contains three classes: children in danger, women in danger, and normal situations. The dataset partition is 70% training set and 30% testing set. The audio files are distributed in two main folders, namely train and test. Each folder is divided into three subfolders, one for each class. The names of those folders are Child (which contains audio files for Children in danger class), Women (which contains audio files for Women in danger class), and Normal (which contains audio files for normal situations). This dataset has 19375 audio files in total, and each file is chopped into 1s. The Children in Danger class has 6488 audio files; among those, 4542 are in the Child subfolder under the Train folder, and 1946 are in the Child subfolder under the Test folder. On the other hand, the Women in Danger class has 6427 audio files; among those, 4498 files are in the Women subfolder under the train folder, and 1929 files are in the Women subfolder under the test folder. Finally, the Normal Situation class has 6460 audio files, among which 4522 are in the Women subfolder under the train folder and 1938 are in the Women subfolder under the test folder.