M.S. Candidate: Atıl İlerialkan
Program: Multimedia Informatics
Date: 27.11.2019 / 15.00
Abstract: Acoustic features extracted from speech are widely used for problems such as biometric speaker identification or first-person activity detection. However, use of speech data raises concerns about privacy due to the explicit availability of the speech content. In this thesis, we propose a method for speech and posture classification using intra-speech breathing sounds.
The acoustical instantaneous side information was extracted from breath instances using the Hilbert-Huang transform. Instantaneous frequency, magnitude, and phase features were extracted using intrinsic mode functions, and different combinations of these were fed into a CNN-RNN network for classification. We also created a publicly available breath dataset, BreathBase, for both our experiments in the thesis and future work. BreathBase contains more than 5000 breath instances detected on the recordings of 20 participants reading pre-prepared random pseudo texts in 5 different postures with 4 different microphones.
Using side information acquired from breath sections of speech, 85% speaker classification and 98% posture classification accuracy is obtained among 20 speakers with this method. The proposed method outperformed various other methods such as support vector machines, long-short term memory and combination of k-nearest neighbor and dynamic time warping techniques.