In this paper, we describe a system for recording of mood diaries in the context of an ambient assisted living and intelligent coaching environment, which ensures privacy by design. The system performs affect recognition in speech features without recording speech content in any form. We demonstrate results of affect recognition models tested on data collected in care-home settings during the SAAM project (Supporting Active Ageing through Multimodal Coaching) using our custom-designed audio collection hardware. The proposed system was trained using Bulgarian speech augmented with training data obtained from comparable mood diaries recorded in Scottish English. A degree of transfer learning of Scottish English speech to Bulgarian speech was demonstrated.
|Publication status||Published - 18 Sep 2022|
|Event||Interspeech 2022 - Incheon, Korea, Democratic People's Republic of|
Duration: 18 Sep 2022 → 22 Sep 2022
Conference number: 23
|Country/Territory||Korea, Democratic People's Republic of|
|Period||18/09/22 → 22/09/22|