FrAUG: A Frame Rate Based Data Augmentation Method for Depression Detection from Speech Signals

Abstract

In this paper, a data augmentation method is proposed for depression detection from speech signals. Samples for data augmentation were created by changing the frame-width and the frame-shift parameters during the feature extraction process. Unlike other data augmentation methods (such as VTLP, pitch perturbation, or speed perturbation), the proposed method does not explicitly change acoustic parameters but rather the time-frequency resolution of frame-level features. The proposed method was evaluated using two different datasets, models, and input acoustic features. For the DAIC-WOZ (English) dataset when using the DepAudioNet model and mel-Spectrograms as input, the proposed method resulted in an improvement of 5.97% (validation) and 25.13% (test) when compared to the baseline. The improvements for the CON-VERGE (Mandarin) dataset when using the x-vector embeddings with CNN as the backend and MFCCs as input features were 9.32% (validation) and 12.99% (test). Baseline systems do not incorporate any data augmentation. Further, the proposed method outperformed commonly used data-augmentation methods such as noise augmentation, VTLP, Speed, and Pitch Perturbation. All improvements were statistically significant.

A Step Towards Preserving Speakers' Identity While Detecting Depression Via Speaker Disentanglement

“…Among others, acoustic features such as x-vectors [18], i-vectors [19] and other speaker embeddings [20] have been shown to be effective in the diagnosis of a speaker's mental state. These features, however, also carry information about a speaker's identity [21] which can be counter-productive to privacy preservation-a key factor in the adoption of digital mental-health screening systems [22].…”