Eliathamby Ambikairajah

Signal Processing in Sequence Analysis: Advances in Eukaryotic Gene Prediction

Genomic sequence processing has been an active area of research for the past two decades and has increasingly attracted the attention of digital signal processing researchers in recent years. A challenging open problem in deoxyribonucleic acid (DNA) sequence analysis is maximizing the prediction accuracy of eukaryotic gene locations and thereby protein coding regions. In this paper, DNA symbolic-to-numeric representations are presented and compared with existing techniques in terms of relative accuracy for the gene and exon prediction problem. Novel signal processing-based gene and exon prediction methods are then evaluated together with existing approaches at a nucleotide level using the Burset/Guigo1996, HMR195, and GENSCAN standard genomic datasets. A new technique for the recognition of acceptor splice sites is then proposed, which combines signal processing-based gene and exon prediction methods with an existing data-driven statistical method. By comparison with the acceptor splice site detection method used in the gene-finding program GENSCAN, the proposed DSP-statistical hybrid technique reveals a consistent reduction in false positives at different levels of sensitivity, averaging a 43% reduction when evaluated on the GENSCAN test set.

Accelerometry Based Classification of Walking Patterns Using Time-frequency Analysis

In this work, 33 dimensional time-frequency domain features were developed and evaluated to detect five different human walking patterns from data acquired using a triaxial accelerometer attached at the waist above the iliac spine. 52 subjects were asked to walk on a flat surface along a corridor, walk up and down a flight of a stairway and walk up and down a constant gradient slope, in an unsupervised manner. Time-frequency domain features of acceleration data in anterior-posterior (AP), medio-lateral (ML) and vertical (VT) direction were developed. The acceleration signal in each direction was decomposed to six detailed signals at different wavelet scales by using the wavelet packet transform. The rms values and standard deviations of the decomposed signals at scales 5 to 2 corresponding to the 0.78-18.75 Hz frequency band were calculated. The energies in the 0.39-18.75 Hz frequency band of acceleration signal in AP, ML and VT directions were also computed. The back-end of the system was a multi-layer perceptron (MLP) Neural Networks (NNs) classifier. Overall classification accuracies of 88.54% and 92.05% were achieved by using a round robin (RR) and random frame selecting (RFS) train-test method respectively for the five walking patterns.

On DNA Numerical Representations for Period-3 Based Exon Prediction

Processing of DNA sequences using traditional digital signal processing methods requires their conversion from a character string into numerical sequences as a first step. Many representations introduced previously assign values to indicate the four DNA nucleotides A, C, G, and T that impose mathematical structures not present in the actual DNA sequence. In this paper, almost all existing methods are compared for the purpose of identifying protein coding regions, using the discrete Fourier transform (DFT) based spectral content measure to exploit period-3 behaviour in the exonic regions for the GENSCAN test set. False positive vs. sensitivity, receiver operating characteristic (ROC) curve and exonic nucleotides detected as false positive results all show that the two newly proposed numerical of DNA representations perform better than the well-known Z-curve, tetrahedron, and Voss representations, with 66-75% less processing. By comparison with Voss representation, the proposed paired numeric method can produce relative improvements of up to 12% in terms of prediction accuracy of exonic nucleotides at a 10% false positive rate using the GENSCAN test set.

An Adapted Gaussian Mixture Model Approach to Accelerometry- Based Movement Classification Using Time-Domain Features

The accurate classification of everyday movements from accelerometry data will provide a significant step towards the development of effective ambulatory monitoring systems for falls detection and prediction. The search continues for optimal front-end processing methods for use in accelerometry systems. Here, we propose a novel set of time domain features, which achieve a mean accuracy of 91.3% in distinguishing between three postures (sitting, standing and lying) and five movements (sit-to-stand, stand-to-sit, lie-to-stand, stand-to-lie and walking). This is a 39.2% relative improvement in error rate over more commonly used frequency based features. A method for adapting Gaussian Mixture Models to compensate for the problem of limited user-specific training data is also proposed and investigated. The method, which uses Bayesian adaptation, was found to improve classification performance for time domain features, offering a mean relative improvement of 20.2% over a non subject-specific system and 4.5% over a system trained using subject specific data only

Speaker verification using sparse representation classification

Sparse representations of signals have received a great deal of attention in recent years, and the sparse representation classifier has very lately appeared in a speaker recognition system. This approach represents the (sparse) GMM mean supervector of an unknown speaker as a linear combination of an over-complete dictionary of GMM supervectors of many speaker models, and ℓ1-norm minimization results in a non-zero coefficient corresponding to the unknown speaker class index. Here this approach is tested on large databases, introducing channel-/session-variability compensation, and fused with a GMM-SVM system. Evaluations on the NIST 2001 SRE and NIST 2006 SRE database show that when the outputs of the MFCC UBM-GMM based classifier (for NIST 2001 SRE) or MFCC GMM-SVM based classifier (for NIST 2006 SRE) are fused with the MFCC GMM-Sparse Representation Classifier (GMM-SRC) based classifier, an absolute gain of 1.27% and 0.25% in EER can be achieved respectively.

Language Identification: A Tutorial

This tutorial presents an overview of the progression of spoken language identification (LID) systems and current developments. The introduction provides a background on automatic language identification systems using syntactic, morphological, and in particular, acoustic, phonetic, phonotactic and prosodic level information. Different frontend features that are used in LID systems are presented. Several normalization and language modelling techniques have also been presented. We also discuss different LID system architectures that embrace a variety of front-ends and back-ends, and configurations such as hierarchical and fusion classifiers. Evaluations of the LID system are presented using NIST language recognition evaluation tasks.

Speech-based cognitive load monitoring system

Monitoring cognitive load is important for the prevention of faulty errors in task-critical operations, and the development of adaptive user interfaces, to maintain productivity and efficiency in work performance. Speech, as an objective and non-intrusive measure, is a suitable method for monitoring cognitive load. Existing approaches for cognitive load monitoring are limited in speaker-dependent recognition and need manually labeled data. We propose a novel automatic, speaker-independent classification approach to monitor, in real-time, the persons cognitive load level by using speech features. In this approach, a Gaussian mixture model (GMM) based classifier is created with unsupervised training. Channel and speaker normalization are deployed for improving robustness. Different delta techniques are investigated for capturing temporal information. And a background model is introduced to reduce the impact of insufficient training data. The final system achieves 71.1% and 77.5% accuracy on two different tasks, each of which has three discrete cognitive load levels. This performance shows a great potential in real-world applications.

Speaker Normalisation for Speech-Based Emotion Detection

The focus of this paper is on speech-based emotion detection utilising only acoustic data, i.e. without using any linguistic or semantic information. However, this approach in general suffers from the fact that acoustic data is speaker-dependent, and can result in inefficient estimation of the statistics modelled by classifiers such as hidden Markov models (HMMs) and Gaussian mixture models (GMMs). We propose the use of speaker-specific feature warping as a means of normalising acoustic features to overcome the problem of speaker dependency. In this paper we compare the performance of a system that uses feature warping to one that does not. The back-end employs an HMM-based classifier that captures the temporal variations of the feature vectors by modelling them as transitions between different states. Evaluations conducted on the LDC emotional prosody speech corpus reveal a relative increase in classification accuracy of up to 20%.

Wideband speech and audio coding using gammatone filter banks

Considerable research attention has been directed towards speech and audio coding algorithms capable of producing high quality coded speech and audio, however few of these use signal representations which account for temporal as well as spectral detail. This paper presents a new technique for 16 kHz wideband speech and audio coding, whereby analysis and synthesis are performed using a linear phase gammatone filter bank. The outputs of these critical band filters are processed to obtain a series of pulse trains that represent neural firing. Auditory masking is then applied to reduce the number of pulses, producing a more compact time-frequency parameterization. The critical band gains and pulse amplitudes and positions are then coded using a combination of non-uniform quantization, arithmetic coding and vector quantization. This coding paradigm produces high quality coded speech and audio, is based upon well-known models of the auditory system, is highly scalable, and has moderate complexity.

Optimizing period-3 methods for eukaryotic gene prediction

In this paper, we firstly investigate the effect of window lengths on selected signal processing-based gene and exon prediction methods. We then optimize these methods to improve their prediction accuracy by employing the best DNA representation, a suitable window length, and boosting the output signals to enhance protein coding and suppress the non-coding regions. It is shown herein that the proposed method outperforms major existing time-domain, frequency- domain, and combined time-frequency approaches. By comparison with the existing DFT-based methods, the proposed method not only requires 50% less processing but also exhibits relative improvements of 53.3%, 46.7%, and 24.2% respectively over spectral content, spectral rotation, and paired and weighted spectral rotation measures in terms of prediction accuracy of exonic nucleotides at a 5% false positive rate using the GENSCAN test set.