Pediatric Obstructive Sleep Apnea-Hypopnea Syndrome (P-OSAHS) leads to cognitive deficits and growth failure if not diagnosed and treated earlier. However, there are significant challenges in diagnosing P-OSAHS using snoring alone, as a proportion of P-OSAHS patients make noisy breathing sounds rather than snoring. In this paper, we introduce the SleepAudio-Child dataset, which includes snoring and breathing sounds relevant to diagnosing OSAHS in children. The dataset comprises 2813 breathing recordings, annotated with 5 labels reflecting all the respiratory events: hypopnea-related breathing, obstructive apnea-related breathing, central apnea-related breathing, mixed apnea-related breathing and normal breathing. The SleepAudio-Child dataset aims to provide comprehensive labeling data for diagnosing OSAHS in children, making it clinically applicable.
| Labels | Wav | Time Waveform | Log Mel Spectrogram |
|---|---|---|---|
| Hy-related |
|
|
|
| OA-related |
|
|
|
| CA-related |
|
|
|
| MA-related |
|
|
|
| Normal |
|
|
| Labels | Wav | Time Waveform | Log Mel Spectrogram |
|---|---|---|---|
| Hy-related |
|
|
|
| OA-related |
|
|
|
| CA-related |
|
|
|
| MA-related |
|
|
|
| Normal |
|
|