Speaker Recognition In Noisy Conditions With Limited Training Data

In this paper we present a novel method for performing speaker recognition with very limited training data and in the presence of background noise. Similarity-based speaker recognition is considered so that speaker models can be created with limited training speech data. The proposed similarity is a form of cosine similarity used as a distance measure between speech feature vectors. Each speech frame is modelled using subband features, and into this framework, multicondition training and optimal feature selection are introduced, making the system capable of performing speaker recognition in the presence of realistic, time-varying noise, which is unknown during training. Speaker identi?cation experiments were carried out using the SPIDRE database. The performance of the proposed new system for noise compensation is compared to that of an oracle model; the speaker identi?cation accuracy for clean speech by the new system trained with limited training data is compared to that of a GMM trained with several minutes of speech. Both comparisons have demonstrated the effectiveness of the new model. Finally, experiments were carried out to test the new model for speaker identi?cation given limited training data and with differing levels and types of realistic background noise. The results have demonstrated the robustness of the new system.

Combining missing-feature theory, speech enhancement, and speaker-dependent/-independent modeling for speech separation

This paper considers the separation and recognition of overlapped speech sentences assuming single-channel observation. A system based on a combination of several different techniques is proposed. The system uses a missing-feature approach for improving crosstalk/noise robustness, a Wiener filter for speech enhancement, hidden Markov models for speech reconstruction, and speaker-dependent/-independent modeling for speaker and speech recognition. We develop the system on the Speech Separation Challenge database, involving a task of separating and recognizing two mixing sentences without assuming advanced knowledge about the identity of the speakers nor about the signal-to-noise ratio. The paper is an extended version of a previous conference paper submitted for the challenge.

Modelling Damage in Thin-Walled Structures [Keynote speaker]

A high-fidelity composite damage model is presented and applied to predict low-velocity impact damage, compression after impact (CAI) strength and crushing of thin-walled composite structures. The simulated results correlated well with experimental testing in terms of overall force-displacement response, damage morphologies and energy dissipation. The predictive power of this model makes it suitable for use as part of a virtual testing methodology, reducing the reliance on physical testing.  

Speaker identification using models for phonemes

Motivation for Speaker recognition work is presented in the first part of the thesis. An exhaustive survey of past work in this field is also presented. A low cost system not including complex computation has been chosen for implementation. Towards achieving this a PC based system is designed and developed. A front end analog to digital convertor (12 bit) is built and interfaced to a PC. Software to control the ADC and to perform various analytical functions including feature vector evaluation is developed. It is shown that a fixed set of phrases incorporating evenly balanced phonemes is aptly suited for the speaker recognition work at hand. A set of phrases are chosen for recognition. Two new methods are adopted for the feature evaluation. Some new measurements involving a symmetry check method for pitch period detection and ACE‘ are used as featured. Arguments are provided to show the need for a new model for speech production. Starting from heuristic, a knowledge based (KB) speech production model is presented. In this model, a KB provides impulses to a voice producing mechanism and constant correction is applied via a feedback path. It is this correction that differs from speaker to speaker. Methods of defining measurable parameters for use as features are described. Algorithms for speaker recognition are developed and implemented. Two methods are presented. The first is based on the model postulated. Here the entropy on the utterance of a phoneme is evaluated. The transitions of voiced regions are used as speaker dependent features. The second method presented uses features found in other works, but evaluated differently. A knock—out scheme is used to provide the weightage values for the selection of features. Results of implementation are presented which show on an average of 80% recognition. It is also shown that if there are long gaps between sessions, the performance deteriorates and is speaker dependent. Cross recognition percentages are also presented and this in the worst case rises to 30% while the best case is 0%. Suggestions for further work are given in the concluding chapter.

Efficient Watermarking Schemes for Speaker Verification Guaranteeing Non-repudiation

Presently different audio watermarking methods are available; most of them inclined towards copyright protection and copy protection. This is the key motive for the notion to develop a speaker verification scheme that guar- antees non-repudiation services and the thesis is its outcome. The research presented in this thesis scrutinizes the field of audio water- marking and the outcome is a speaker verification scheme that is proficient in addressing issues allied to non-repudiation to a great extent. This work aimed in developing novel audio watermarking schemes utilizing the fun- damental ideas of Fast-Fourier Transform (FFT) or Fast Walsh-Hadamard Transform (FWHT). The Mel-Frequency Cepstral Coefficients (MFCC) the best parametric representation of the acoustic signals along with few other key acoustic characteristics is employed in crafting of new schemes. The au- dio watermark created is entirely dependent to the acoustic features, hence named as FeatureMark and is crucial in this work. In any watermarking scheme, the quality of the extracted watermark de- pends exclusively on the pre-processing action and in this work framing and windowing techniques are involved. The theme non-repudiation provides immense significance in the audio watermarking schemes proposed in this work. Modification of the signal spectrum is achieved in a variety of ways by selecting appropriate FFT/FWHT coefficients and the watermarking schemes were evaluated for imperceptibility, robustness and capacity char- acteristics. The proposed schemes are unequivocally effective in terms of maintaining the sound quality, retrieving the embedded FeatureMark and in terms of the capacity to hold the mark bits. Robust nature of these marking schemes is achieved with the help of syn- chronization codes such as Barker Code with FFT based FeatureMarking scheme and Walsh Code with FWHT based FeatureMarking scheme. An- other important feature associated with this scheme is the employment of an encryption scheme towards the preparation of its FeatureMark that scrambles the signal features that helps to keep the signal features unreve- laed. A comparative study with the existing watermarking schemes and the ex- periments to evaluate imperceptibility, robustness and capacity tests guar- antee that the proposed schemes can be baselined as efficient audio water- marking schemes. The four new digital audio watermarking algorithms in terms of their performance are remarkable thereby opening more opportu- nities for further research.

Mediating between hearer's and speaker's views in the generation of adaptive explanations

In this article, we examine the case of a system that cooperates with a “direct” user to plan an activity that some “indirect” user, not interacting with the system, should perform. The specific application we consider is the prescription of drugs. In this case, the direct user is the prescriber and the indirect user is the person who is responsible for performing the therapy. Relevant characteristics of the two users are represented in two user models. Explanation strategies are represented in planning operators whose preconditions encode the cognitive state of the indirect user; this allows tailoring the message to the indirect user's characteristics. Expansion of optional subgoals and selection among candidate operators is made by applying decision criteria represented as metarules, that negotiate between direct and indirect users' views also taking into account the context where explanation is provided. After the message has been generated, the direct user may ask to add or remove some items, or change the message style. The system defends the indirect user's needs as far as possible by mentioning the rationale behind the generated message. If needed, the plan is repaired and the direct user model is revised accordingly, so that the system learns progressively to generate messages suited to the preferences of people with whom it interacts.