In natural listening environments, speech signals are easily distorted by variousacoustic interference, which reduces the speech quality and intelligibility of human listening; meanwhile, it makes difficult for many speech-related applications, such as automatic speech recognition (ASR). Thus, many speech enhancement (SE) algorithms have been developed in the past decades. However, most current SE algorithms are difficult to capture underlying speech information (e.g., phoneme) in the SE process. This causes it to be challenging to know what specific information is lost or interfered with in the SE process, which limits the application of enhanced speech. For instance, some SE algorithms aimed to improve human listening usually damage the ASR system. The objective of this dissertation is to develop SE algorithms that have the potential to capture various underlying speech representations (information) and improve the quality and intelligibility of noisy speech. This ...

One desideratum in designing cognitive robots is autonomous learning of communication skills, just like humans. The primary step towards this goal is vocabulary acquisition. Being different from the training procedures of the state-of-the-art automatic speech recognition (ASR) systems, vocabulary acquisition cannot rely on prior knowledge of language in the same way. Like what infants do, the acquisition process should be data-driven with multi-level abstraction and coupled with multi-modal inputs. To avoid lengthy training efforts in a word-by-word interactive learning process, a clever learning agent should be able to acquire vocabularies from continuous speech automatically. The work presented in this thesis is entitled \emph{Constrained Non-negative Matrix Factorization for Vocabulary Acquisition from Continuous Speech}. Enlightened by the extensively studied techniques in ASR, we design computational models to discover and represent vocabularies from continuous speech with little prior knowledge of the language to ...

This thesis presents a novel approach to speech enhancement by exploiting the bimodality of speech production and the correlation that exists between audio and visual speech information. An analysis into the correlation of a range of audio and visual features reveals significant correlation to exist between visual speech features and audio filterbank features. The amount of correlation was also found to be greater when the correlation is analysed with individual phonemes rather than across all phonemes. This led to building a Gaussian Mixture Model (GMM) that is capable of estimating filterbank features from visual features. Phoneme-specific GMMs gave lower filterbank estimation errors and a phoneme transcription is decoded using audio-visual Hidden Markov Model (HMM). Clean filterbank estimates along with mean noise estimates were then utilised to construct visually-derived Wiener filters that are able to enhance noisy speech. The mean noise ...

This dissertation introduces interactive designs in the context of statistical parametric synthesis. The objective is to develop methods and designs that enrich the Human-Computer Interaction by enabling computers (or other devices) to have more expressive and adjustable voices. First, we tackle the problem of interactive controls and present a novel method for performative HMM-based synthesis (pHTS). Second, we apply interpolation methods, initially developed for the traditional HMM-based speech synthesis system, in the interactive framework of pHTS. Third, we integrate articulatory control in our interactive approach. Fourth, we present a collection of interactive applications based on our work. Finally, we unify our research into an open source library, Mage. To our current knowledge Mage is the first system for interactive programming of HMM-based synthesis that allows realtime manipulation of all speech production levels. It has been used also in cases that ...

The interest in the field of speech enhancement emerges from the increased usage of digital speech processing applications like mobile telephony, digital hearing aids and human-machine communication systems in our daily life. The trend to make these applications mobile increases the variety of potential sources for quality degradation. Speech enhancement methods can be used to increase the quality of these speech processing devices and make them more robust under noisy conditions. The name "speech enhancement" refers to a large group of methods that are all meant to improve certain quality aspects of these devices. Examples of speech enhancement algorithms are echo control, bandwidth extension, packet loss concealment and noise reduction. In this thesis we focus on single-microphone additive noise reduction and aim at methods that work in the discrete Fourier transform (DFT) domain. The main objective of the presented research ...

According to the Parkinson’s Foundation, more than 10 million people world- wide suffer from Parkinson’s disease (PD). The common symptoms are tremor, muscle rigidity and slowness of movement. There is no cure available cur- rently, 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 envi- ronments, such as noise, reverberation and nonlinear distortion, which are common in uncontrolled settings. In this thesis, we focus on deriving speech modeling and robust estima- tion 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 research in this thesis addresses the problem of automatic speech recognition in noisy environments. Automatic speech recognition systems obtain acceptable performances in noise free conditions but these performances degrade dramatically in presence of additive noise. This is mainly due to the mismatch between the training and the noisy operating conditions. In the time-frequency representation of the noisy speech signal, some of the clean speech features are masked by noise. In this case the clean speech features cannot be correctly estimated from the noisy speech and therefore they are considered as missing or unreliable. In order to improve the performance of speech recognition systems in additive noise conditions, special attention should be paid to the problems of detection and compensation of these unreliable features. This thesis is concerned with the problem of missing features applied to automatic speaker-independent speech recognition. ...

This thesis presents novel exemplar-based texture synthesis methods for image prediction (i.e., predictive coding) and image inpainting problems. The main contributions of this study can also be seen as extensions to simple template matching, however the texture synthesis problem here is well-formulated in an optimization framework with different constraints. The image prediction problem has first been put into sparse representations framework by approximating the template with a sparsity constraint. The proposed sparse prediction method with locally and adaptive dictionaries has been shown to give better performance when compared to static waveform (such as DCT) dictionaries, and also to the template matching method. The image prediction problem has later been placed into an online dictionary learning framework by adapting conventional dictionary learning approaches for image prediction. The experimental observations show a better performance when compared to H.264/AVC intra and sparse prediction. ...

Never before has speech been captured as often by electronic devices equipped with one or multiple microphones, serving a variety of applications. It is the key aspect in digital telephony, hearing devices, and voice-driven human-to-machine interaction. When speech is recorded, the microphones also capture a variety of further, undesired sound components due to adverse acoustic conditions. Interfering speech, background noise and reverberation, i.e. the persistence of sound in a room after excitation caused by a multitude of reflections on the room enclosure, are detrimental to the quality and intelligibility of target speech as well as the performance of automatic speech recognition. Hence, speech enhancement aiming at estimating the early target-speech component, which contains the direct component and early reflections, is crucial to nearly all speech-related applications presently available. In this thesis, we compare, propose and evaluate existing and novel approaches ...

In modern devices which provide hands-free speech capturing functionality, such as hands-free communication kits and voice-controlled devices, the received speech signal at the microphones is corrupted by background noise, interfering speech signals, and room reverberation. In many practical situations, the microphones are not necessarily located near the desired source, and hence, the ratio of the desired speech power to the power of the background noise, the interfering speech, and the reverberation at the microphones can be very low, often around or even below 0 dB. In such situations, the comfort of human-to-human communication, as well as the accuracy of automatic speech recognisers for voice-controlled applications can be signi cantly degraded. Therefore, e ffective speech enhancement algorithms are required to process the microphone signals before transmitting them to the far-end side for communication, or before feeding them into a speech recognition ...

Human action recognition problem is one of the most challenging problems in the computer vision domain, and plays an emerging role in various fields of study. In this thesis, we investigate structured and sequential representations of spatio-temporal data for recognizing human actions and for measuring action performance quality. In video sequences, we characterize each action with a graphical structure of its spatio-temporal interest points and each such interest point is qualified by its cuboid descriptors. In the case of depth data, an action is represented by the sequence of skeleton joints. Given such descriptors, we solve the human action recognition problem through a hyper-graph matching formulation. As is known, hyper-graph matching problem is NP-complete. We simplify the problem in two stages to enable a fast solution: In the first stage, we take into consideration the physical constraints such as time ...

This dissertation's primary goal was to systematically study human activity recognition and enhance its performance by advancing human activities' sequential modeling based on HMM-based machine learning. Driven by these purposes, this dissertation has the following major contributions: The proposal of our HAR research pipeline that guides the building of a robust wearable end-to-end HAR system and the implementation of the recording and recognition software Activity Signal Kit (ASK) according to the pipeline; Collecting several datasets of multimodal biosignals from over 25 subjects using the self-implemented ASK software and implementing an easy mechanism to segment and annotate the data; The comprehensive research on the offline HAR system based on the recorded datasets and the implementation of an end-to-end real-time HAR system; A novel activity modeling method for HAR, which partitions the human activity into a sequence of shared, meaningful, and activity ...

Hyperspectral (HS) imaging, which consists of acquiring a same scene in several hundreds of contiguous spectral bands (a three dimensional data cube), has opened a new range of relevant applications, such as target detection [MS02], classification [C.-03] and spectral unmixing [BDPD+12]. However, while HS sensors provide abundant spectral information, their spatial resolution is generally more limited. Thus, fusing the HS image with other highly resolved images of the same scene, such as multispectral (MS) or panchromatic (PAN) images is an interesting problem. The problem of fusing a high spectral and low spatial resolution image with an auxiliary image of higher spatial but lower spectral resolution, also known as multi-resolution image fusion, has been explored for many years [AMV+11]. From an application point of view, this problem is also important as motivated by recent national programs, e.g., the Japanese next-generation space-borne ...

This thesis deals with minimum mean square error (MMSE) speech enhancement schemes in the short-time Fourier transform (STFT) domain with a focus on statistical models for speech and corresponding estimators. MMSE speech enhancement approaches taking speech presence uncertainty (SPU) into account usually consist of a common MMSE estimator for speech and an a posteriori speech presence probability (SPP) estimator. It is shown that both estimators should be based on the same statistical speech model, as they are in the same estimation framework and assume the same a priori knowledge. In order to give a synopsis of consistent MMSE estimation under SPU, typical common MMSE estimators and a posteriori SPP estimators are recapitulated. Furthermore, a new specific a posteriori SPP estimator is derived based on a novel statistical model for speech. Then, a synopsis of approaches to consistent MMSE estimation under ...

The topic of this thesis is speech enhancement with a focus on models of voiced speech. Speech is divided into two subcategories dependent on the characteristics of the signal. One part is the voiced speech, the other is the unvoiced. In this thesis, we primarily focus on the voiced speech parts and utilise the structure of the signal in relation to speech enhancement. The basis for the models is the harmonic model which is a very often used model for voiced speech because it describes periodic signals perfectly. First, we consider the problem of non-stationarity in the speech signal. The speech signal changes its characteristics continuously over time whereas most speech analysis and enhancement methods assume stationarity within 20-30 ms. We propose to change the model to allow the fundamental frequency to vary linearly over time by introducing a chirp ...