The defense will take place on January 21st 2019 at 1pm.
Venue: Rendsburggade 14, room: 3.563 and 3.565, 9000 Aalborg.
The dissertation is a result of a PhD study at the Department of Architecture, Design and Media Technology, Aalborg University, Denmark.
After the defense, the Department of Architecture, Design and Media Technology will host a reception in room 5.355 (lunch area, level 3).
Questions can be directed to secretary Kristina Wagner Røjen (+4599409926/ firstname.lastname@example.org).
Summary of the PhD research:
According to the Parkinson's Foundation, more than 10 million people worldwide suffer from Parkinson's disease (PD). The common symptoms are tremor, muscle rigidity and slowness of movement. There is no cure available currently, but clinical intervention can help alleviate the symptoms significantly. Recently, it has been found that PD can be detected and telemonitored by voice signals, such as sustained phonation /a/. However, the voiced-based PD detector suffers from severe performance degradation in adverse environments, such as noise, reverberation and nonlinear distortion, which are common in uncontrolled settings.
In this thesis, we focus on deriving speech modeling and robust estimation algorithms capable of improving the PD detection accuracy in adverse environments. Robust estimation algorithms using parametric modeling of voice signals are proposed. We present both segment-wise and sample-wise robust pitch tracking algorithms using the harmonic model. The first order Markov chain is used to impose smoothness prior for the pitch. In segment-wise pitch tracking, we have proposed a method to track the pitch, harmonic order and voicing state jointly based on Bayesian tracking framework. In sample-wise pitch tracking, to deal with colored noise, the noise is modeled as time-varying autoregressive process. The proposed algorithms are compared with the state-of-the-art pitch estimation algorithms and are evaluated on the Parkinson's disease database. Apart from extracting pitch information, we have also looked into the problem of autoregressive moving average (ARMA) modeling of voiced speech and its parameters estimation. Due to the sparse nature of the excitation signal for the voiced speech, both least 1-norm criterion and sparse Bayesian learning are applied to improve the ARMA coefficients estimation. The proposed ARMA estimation methods are shown to perform better than the least squares-based method in terms of spectral distortion. We have also proposed a dictionary-based speech enhancement algorithm using non-negative matrix factorization, where the dictionary items for both speech and noise are parameterized by AR coefficients. Finally, we investigated on the performance of a vast number of speech enhancement and dereverberation algorithms for diagnosis of PD with degraded speech signals.