Improving out-domain PLDA speaker verification using unsupervised inter-dataset variability compensation approach

Experimental studies have found that when the state-of-the-art probabilistic linear discriminant analysis (PLDA) speaker verification systems are trained using out-domain data, it significantly affects speaker verification performance due to the mismatch between development data and evaluation data. To overcome this problem we propose a novel unsupervised inter dataset variability (IDV) compensation approach to compensate the dataset mismatch. IDV-compensated PLDA system achieves over 10% relative improvement in EER values over out-domain PLDA system by effectively compensating the mismatch between in-domain and out-domain data.

Improving PLDA speaker verification using WMFD and linear-weighted approaches in limited microphone data conditions

This paper proposes the addition of a weighted median Fisher discriminator (WMFD) projection prior to length-normalised Gaussian probabilistic linear discriminant analysis (GPLDA) modelling in order to compensate the additional session variation. In limited microphone data conditions, a linear-weighted approach is introduced to increase the influence of microphone speech dataset. The linear-weighted WMFD-projected GPLDA system shows improvements in EER and DCF values over the pooled LDA- and WMFD-projected GPLDA systems in inter-view-interview condition as WMFD projection extracts more speaker discriminant information with limited number of sessions/ speaker data, and linear-weighted GPLDA approach estimates reliable model parameters with limited microphone data.

Dataset-invariant covariance normalization for out-domain PLDA speaker verification

In this paper we introduce a novel domain-invariant covariance normalization (DICN) technique to relocate both in-domain and out-domain i-vectors into a third dataset-invariant space, providing an improvement for out-domain PLDA speaker verification with a very small number of unlabelled in-domain adaptation i-vectors. By capturing the dataset variance from a global mean using both development out-domain i-vectors and limited unlabelled in-domain i-vectors, we could obtain domain- invariant representations of PLDA training data. The DICN- compensated out-domain PLDA system is shown to perform as well as in-domain PLDA training with as few as 500 unlabelled in-domain i-vectors for NIST-2010 SRE and 2000 unlabelled in-domain i-vectors for NIST-2008 SRE, and considerable relative improvement over both out-domain and in-domain PLDA development if more are available.

A cluster-voting approach for speaker diarization and linking of Australian broadcast news recordings

We present a clustering-only approach to the problem of speaker diarization to eliminate the need for the commonly employed and computationally expensive Viterbi segmentation and realignment stage. We use multiple linear segmentations of a recording and carry out complete-linkage clustering within each segmentation scenario to obtain a set of clustering decisions for each case. We then collect all clustering decisions, across all cases, to compute a pairwise vote between the segments and conduct complete-linkage clustering to cluster them at a resolution equal to the minimum segment length used in the linear segmentations. We use our proposed cluster-voting approach to carry out speaker diarization and linking across the SAIVT-BNEWS corpus of Australian broadcast news data. We compare our technique to an equivalent baseline system with Viterbi realignment and show that our approach can outperform the baseline technique with respect to the diarization error rate (DER) and attribution error rate (AER).

The QUT-NOISE-SRE protocol for the evaluation of noisy speaker recognition

The QUT-NOISE-SRE protocol is designed to mix the large QUT-NOISE database, consisting of over 10 hours of back- ground noise, collected across 10 unique locations covering 5 common noise scenarios, with commonly used speaker recognition datasets such as Switchboard, Mixer and the speaker recognition evaluation (SRE) datasets provided by NIST. By allowing common, clean, speech corpora to be mixed with a wide variety of noise conditions, environmental reverberant responses, and signal-to-noise ratios, this protocol provides a solid basis for the development, evaluation and benchmarking of robust speaker recognition algorithms, and is freely available to download alongside the QUT-NOISE database. In this work, we use the QUT-NOISE-SRE protocol to evaluate a state-of-the-art PLDA i-vector speaker recognition system, demonstrating the importance of designing voice-activity-detection front-ends specifically for speaker recognition, rather than aiming for perfect coherence with the true speech/non-speech boundaries.

Use of 3D printed materials as tissue-equivalent phantoms

This study used the specific example of 3D printing with acrylonitrile butadiene styrene (ABS) as a means to investigate the potential usefulness of benchtop rapid prototyping as a technique for producing patient specific phantoms for radiotherapy dosimetry. Three small cylinders and one model of a human lung were produced via in-house 3D printing with ABS, using 90%, 50%, 30% and 10% ABS infill densities. These phantom samples were evaluated in terms of their geometric accuracy, tissue equivalence and radiation hardness, when irradiated using a range of clinical radiotherapy beams. The measured dimensions of the small cylindrical phantoms all matched their planned dimensions, within 1mm. The lung phantom was less accurately matched to the lung geometry on which it was based, due to simplifications introduced during the phantom design process. The mass densities, electron densities and linear attenuation coefficients identified using CT data, as well as the results of film measurements made using megavoltage photon and electron beams, indicated that phantoms printed with ABS, using infill densities of 30% or more, are potentially useful as lung- and tissue-equivalent phantoms for patient-specific radiotherapy dosimetry. All cylindrical 3D printed phantom samples were found to be unaffected by prolonged radiation and to accurately match their design specifications. However, care should be taken to avoid oversimplifying anatomical structures when printing more complex phantoms.

A comparative study of speaker adaptation methods

For the problem of speaker adaptation in speech recognition, the performance depends on the availability of adaptation data. In this paper, we have compared several existing speaker adaptation methods, viz. maximum likelihood linear regression (MLLR), eigenvoice (EV), eigenspace-based MLLR (EMLLR), segmental eigenvoice (SEV) and hierarchical eigenvoice (HEV) based methods. We also develop a new method by modifying the existing HEV method for achieving further performance improvement in a limited available data scenario. In the sense of availability of adaptation data, the new modified HEV (MHEV) method is shown to perform better than all the existing methods throughout the range of operation except the case of MLLR at the availability of more adaptation data.

Hugo Steiner-Prag approaching speaker's lectern at the International Book Arts Exhibition; Leipzig.

Digital Image

Implementation of placement and routing algorithms for computer aided design of printed circuit boards

The Printed Circuit Board (PCB) layout design is one of the most important and time consuming phases during equipment design process in all electronic industries. This paper is concerned with the development and implementation of a computer aided PCB design package. A set of programs which operate on a description of the circuit supplied by the user in the form of a data file and subsequently design the layout of a double-sided PCB has been developed. The algorithms used for the design of the PCB optimise the board area and the length of copper tracks used for the interconnections. The output of the package is the layout drawing of the PCB, drawn on a CALCOMP hard copy plotter and a Tektronix 4012 storage graphics display terminal. The routing density (the board area required for one component) achieved by this package is typically 0.8 sq. inch per IC. The package is implemented on a DEC 1090 system in Pascal and FORTRAN and SIGN(1) graphics package is used for display generation.

Deploying speaker verification techniques to vision-based RPA detect and avoid

Detect and Avoid (DAA) technology is widely acknowledged as a critical enabler for unsegregated Remote Piloted Aircraft (RPA) operations, particularly Beyond Visual Line of Sight (BVLOS). Image-based DAA, in the visible spectrum, is a promising technological option for addressing the challenges DAA presents. Two impediments to progress for this approach are the scarcity of available video footage to train and test algorithms, in conjunction with testing regimes and specifications which facilitate repeatable, statistically valid, performance assessment. This paper includes three key contributions undertaken to address these impediments. In the first instance, we detail our progress towards the creation of a large hybrid collision and near-collision encounter database. Second, we explore the suitability of techniques employed by the biometric research community (Speaker Verification and Language Identification), for DAA performance optimisation and assessment. These techniques include Detection Error Trade-off (DET) curves, Equal Error Rates (EER), and the Detection Cost Function (DCF). Finally, the hybrid database and the speech-based techniques are combined and employed in the assessment of a contemporary, image based DAA system. This system includes stabilisation, morphological filtering and a Hidden Markov Model (HMM) temporal filter.

Hyvinvointivaltion pelastamisella huonosti voivia eläkeläisiä? – mediakeskustelua vuoden 2009 eläkeiän nostopäätöksestä

This thesis examines the media debate on pensions. The case analysed in the thesis is the debate that sharpened after the Finnish government made the decision to raise the retirement age. The analysed data consists of articles published in the printed media during one month after the decision was made on 24th of February in 2009. The aim of the study is to describe how the decision is argued about by different positions of speakers and how retirement is justified from the perspective of the individual. Furthermore, the purpose is to discover different ways of discussing the pensioner. The theoretical frame for this study is social constructivism, which understands reality as socially constructed with language. From this perspective, media texts can be seen as one form of shaping reality. The data is analysed by using different methods. Thematisation is used to discover the key topics, and quantification is used to examine the prevalence of different arguments. The method in which the speaker’s ways of speaking is analysed in different participant categories I call “a speaker position analysis”. The debate around the decision to raise the retirement age highlight the power struggle both between the government and the opposition as well as the government and employee unions. One thing all discussants agree is the need to raise the retirement age. From the individual's perspective, retirement is justified mostly with hard working conditions and inadequacy of health. The pensioner's image is appearing gloomy in most discourses. Prevailing discourses are seeing a pensioner either sick and tired or someone who is not good for work and has lost his dignity. The debate around the decision is intertwined around the concepts of welfare state and individual's well-being. In the postmodern society, human preferences are individualised. Welfare state means different things to different people, as well as the individual's subjective perception of well-being is unique. These two aspects are the ones which raise the tension in the analysed media debate.

Design of speaker identification schemes for large number of speakers (A)

Design of speaker identification schemes for a small number of speakers (around 10) with a high degree of accuracy in controlled environment is a practical proposition today. When the number of speakers is large (say 50–100), many of these schemes cannot be directly extended, as both recognition error and computation time increase monotonically with population size. The feature selection problem is also complex for such schemes. Though there were earlier attempts to rank order features based on statistical distance measures, it has been observed only recently that the best two independent measurements are not the same as the combination in two's for pattern classification. We propose here a systematic approach to the problem using the decision tree or hierarchical classifier with the following objectives: (1) Design of optimal policy at each node of the tree given the tree structure i.e., the tree skeleton and the features to be used at each node. (2) Determination of the optimal feature measurement and decision policy given only the tree skeleton. Applicability of optimization procedures such as dynamic programming in the design of such trees is studied. The experimental results deal with the design of a 50 speaker identification scheme based on this approach.

Machine recognition of printed Kannada text

This paper presents the design of a full fledged OCR system for printed Kannada text. The machine recognition of Kannada characters is difficult due to similarity in the shapes of different characters, script complexity and non-uniqueness in the representation of diacritics. The document image is subject to line segmentation, word segmentation and zone detection. From the zonal information, base characters, vowel modifiers and consonant conjucts are separated. Knowledge based approach is employed for recognizing the base characters. Various features are employed for recognising the characters. These include the coefficients of the Discrete Cosine Transform, Discrete Wavelet Transform and Karhunen-Louve Transform. These features are fed to different classifiers. Structural features are used in the subsequent levels to discriminate confused characters. Use of structural features, increases recognition rate from 93% to 98%. Apart from the classical pattern classification technique of nearest neighbour, Artificial Neural Network (ANN) based classifiers like Back Propogation and Radial Basis Function (RBF) Networks have also been studied. The ANN classifiers are trained in supervised mode using the transform features. Highest recognition rate of 99% is obtained with RBF using second level approximation coefficients of Haar wavelets as the features on presegmented base characters.