DOCSLIB.ORG

Texture Classification Using Spectral Entropy of Acoustic Signal

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Texture Classification Using Spectral Entropy of Acoustic Signal entropy Article Texture Classification Using Spectral Entropy of Acoustic Signal Generated by a Human Echolocator Raja Syamsul Azmir Raja Abdullah 1,*, Nur Luqman Saleh 1, Sharifah Mumtazah Syed Abdul Rahman 1 , Nur Syazmira Zamri 1 and Nur Emileen Abdul Rashid 2 1 Wireless and Photonic Network Research Centre (WiPNET), Faculty of Engineering, University Putra Malaysia (UPM), Serdang 43400, Malaysia; [email protected] (N.L.S.); [email protected] (S.M.S.A.R.); [email protected] (N.S.Z.) 2 Microwave Research Institute, UniversitiTeknologi MARA (UiTM), Shah Alam 40450, Selangor, Malaysia; [email protected] * Correspondence: [email protected]; Tel.: +603-976-943-47 Received: 12 August 2019; Accepted: 29 September 2019; Published: 2 October 2019 Abstract: Human echolocation is a biological process wherein the human emits a punctuated acoustic signal, and the ear analyzes the echo in order to perceive the surroundings. The peculiar acoustic signal is normally produced by clicking inside the mouth. This paper utilized this unique acoustic signal from a human echolocator as a source of transmitted signal in a synthetic human echolocation technique. Thus, the aim of the paper was to extract information from the echo signal and develop a classification scheme to identify signals reflected from different textures at various distance. The scheme was based on spectral entropy extracted from Mel-scale filtering output in the Mel-frequency cepstrum coefficient of a reflected echo signal. The classification process involved data mining, features extraction, clustering, and classifier validation. The reflected echo signals were obtained via an experimental setup resembling a human echolocation scenario, configured for synthetic data collection. Unlike in typical speech signals, extracted entropy from the formant characteristics was likely not visible for the human mouth-click signals. Instead, multiple peak spectral features derived from the synthesis signal of the mouth-click were assumed as the entropy obtained from the Mel-scale filtering output. To realize the classification process, K-means clustering and K-nearest neighbor processes were employed. Moreover, the impacts of sound propagation toward the extracted spectral entropy used in the classification outcome were also investigated. The outcomes of the classifier performance herein indicated that spectral entropy is essential for human echolocation. Keywords: spectral entropy; acoustic signal; human echolocation; classification; MFCC 1. Introduction The phrase “echolocation” was initially used by Griffin to describe bats’ ability to safely navigate and locate their prey using ultrasound call signals [1]. What is less known is that a group of people (often blind people) known as human echolocators have adapted to visualize their surrounding using a similar concept. Human echolocation is the ability to perceive one’s surroundings by listening to the echoes of the active emission of sound signals reflected from an obstacle. Recent studies have reported that these people are able to visualize their surrounding by “seeing through sound”. They exhibit exceptional performance in defining their space, even able to accurately discriminate the profile of objects. This ability has elicited inquiries among scholars on how the space and objects can be recognized using mouth-click signals. Although a series of studies have focused on this question, most works have focused on the perceptual concept rather than technical explanations. It is known that Entropy 2019, 21, 963; doi:10.3390/e21100963 www.mdpi.com/journal/entropy Entropy 2019, 21, 963 2 of 20 Entropy 2019, 21, 963 2 of 20 orderhuman to translate echolocators meaningful depend cues on theirfrom auditorymouth-click systems signals in and order turn to them translate into meaningfulvisual perception, cues from as illustratedmouth-click in Figure signals 1 and[1]. turn them into visual perception, as illustrated in Figure1[1]. FigureFigure 1. 1.TheThe human human echolocation echolocation concept. concept. It Itis isessential essential for for humans humans to to recognize recognize the the ac acousticoustic signals signals entering entering their their ear, ear, mainly mainly for for communicationcommunication and and recognition. recognition. For For comparison, comparison, radar radar and and sonar sonar are are a good a good example example of of man-made man-made sensorssensors which which benefit benefit fromfrom thesethese classifications; classifications; the th illuminatione illumination of a targetof a target is associated is associated with detection with detectionand classification and classification schemes [schemes2–4]. This [2–4]. meaningful This meaningful information information is useful in is distinguishing useful in distinguishing the profile of thethe profile detected of the target, detected and helpstarget, to and minimize helps to poor minimize and false poor detection and false events. detection For humanevents. echolocation,For human echolocation,how people how recognize people spaces recognize and objects spaces using and objects mouth-click using signalsmouth-click has still signals not been has clearlystill not verified. been clearlyIn addition, verified. no In studies addition, to date no have studies reported to date a technicalhave reported classification a technical process classification of human mouth-clicks.process of humanMoreover, mouth-clicks. human mouth-clicks Moreover, dohuma notn inherit mouth-clicks formant propertiesdo not inherit like thoseformant in typical properties speech, like which those have in typicalbeen provenspeech, to which be strong have features been proven in speech to recognition.be strong features Instead, in the speech multiple recognition. frequency Instead, components the multiple(spectral frequency entropy) foundcomponents in the signal(spectral serve entropy) as features found for thein the classification signal serve process. as features This gap for should the classificationbe investigated process. to assure This continuitygap should and be to investigated utilize the full to potentialassure continuity of human and echolocation. to utilize the We full were potentialthus motivated of human to analyzeecholocation. and design We were this thus classification motivated process. to analyze and design this classification process.Studies related to the human auditory system became the primary references in this paper for the classificationStudies related process to the of human human mouth-clicks. auditory system We be hereincame propose the primary a classification references schemein this paper for human for themouth-clicks classification using process experimental of human data mouth-clicks. by utilizing We a human herein auditory propose model a classification (Mel-frequency scheme cepstral for humancoeffi cientmouth-clicks (MFCC) using framework experime processing).ntal data by By utilizing understanding a huma hown auditory human model echolocators (Mel-frequency carry out cepstralecholocation coefficient (utilizing (MFCC) mouth-click), framework a newprocessi dimensionng). By couldunderstanding be opened how in the human design echolocators of man-made carrysensors out (especiallyecholocation for (utilizing radar and mouth-click), sonar) in the neara new future. dimension In this could paper, be we opened have not in madethe design any claim of man-madethat the classification sensors (especially schemes for described radar and closely sonar) replicate in the the near strategy future. used Infor this human paper, echolocation; we have not we madepresent any the claim best intuitivethat the approachedclassification for schemes decision-making described based closely on thereplicate credible the knowledge strategy used of human for humanecholocation echolocation; techniques we present and the humanthe best hearing intuitiv process.e approached for decision-making based on the credibleHence, knowledge this study of human aimed echolocation to investigate techniques the characteristics and the ofhuman the echo hearing signal process. (acoustic mouth click) reflectedHence, from this distudyfferent aimed textures to atinvestigate various distances. the characteristics We developed of the a classificationecho signal (acoustic scheme to mouth classify click)textures reflected based from on thedifferent spectral textures entropy at various of the echo distances. signal We obtained developed from a Mel-scale classification filtering scheme outputs to classify(Mel-spectrum), textures based which on was the incorporated spectral entropy into the of MFCCthe echo framework. signal obtained The classification from Mel-scale tasks filtering included outputsdistinguishing (Mel-spectrum), hard, medium, which andwas softincorporated textures, and into grouping the MFCC them framework. into their The respective classification cluster tasks region includedsources distinguishing using the reflected hard, echo medium, signal and at di softfferent textures, distances. and Thegrouping classification them into routine their wasrespective realized clusterusing region K-means sources process using and the was refl validatedected echo using signal K-nearest at different neighbors distances. (K-NN). The Theclassification paper is structured routine wasas follows:realized Sectionusing K-means2 provides process a description and was of validated the characteristics using K-nearest of the human neighbors mouth-click (K-NN). signal, The paperand presentsis structured a brief as explanation follows:
Load more

Recommended publications
  • A Thesis Entitled Nocturnal Bird Call Recognition System for Wind Farm
    A Thesis entitled Nocturnal Bird Call Recognition System for Wind Farm Applications by Selin A. Bastas Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Master of Science Degree in Electrical Engineering _______________________________________ Dr. Mohsin M. Jamali, Committee Chair _______________________________________ Dr. Junghwan Kim, Committee Member _______________________________________ Dr. Sonmez Sahutoglu, Committee Member _______________________________________ Dr. Patricia R. Komuniecki, Dean College of Graduate Studies The University of Toledo December 2011 Copyright 2011, Selin A. Bastas. This document is copyrighted material. Under copyright law, no parts of this document may be reproduced without the expressed permission of the author. An Abstract of Nocturnal Bird Call Recognition System for Wind Farm Applications by Selin A. Bastas Submitted to the Graduate Faculty as partial fulfillment of the requirements for the Master of Science Degree in Electrical Engineering The University of Toledo December 2011 Interaction of birds with wind turbines has become an important public policy issue. Acoustic monitoring of birds in the vicinity of wind turbines can address this important public policy issue. The identification of nocturnal bird flight calls is also important for various applications such as ornithological studies and acoustic monitoring to prevent the negative effects of wind farms, human made structures and devices on birds. Wind turbines may have negative impact on bird population. Therefore, the development of an acoustic monitoring system is critical for the study of bird behavior. This work can be employed by wildlife biologist for developing mitigation techniques for both on- shore/off-shore wind farm applications and to address bird strike issues at airports.
    [Show full text]
  • The Articulatory and Acoustic Characteristics of Polish Sibilants and Their Consequences for Diachronic Change
    The articulatory and acoustic characteristics of Polish sibilants and their consequences for diachronic change Veronique´ Bukmaier & Jonathan Harrington Institute of Phonetics and Speech Processing, University of Munich, Germany [email protected] [email protected] The study is concerned with the relative synchronic stability of three contrastive sibilant fricatives /s ʂɕ/ in Polish. Tongue movement data were collected from nine first-language Polish speakers producing symmetrical real and non-word CVCV sequences in three vowel contexts. A Gaussian model was used to classify the sibilants from spectral information in the noise and from formant frequencies at vowel onset. The physiological analysis showed an almost complete separation between /s ʂɕ/ on tongue-tip parameters. The acoustic analysis showed that the greater energy at higher frequencies distinguished /s/ in the fricative noise from the other two sibilant categories. The most salient information at vowel onset was for /ɕ/, which also had a strong palatalizing effect on the following vowel. Whereas either the noise or vowel onset was largely sufficient for the identification of /s ɕ/ respectively, both sets of cues were necessary to separate /ʂ/ from /s ɕ/. The greater synchronic instability of /ʂ/ may derive from its high articulatory complexity coupled with its comparatively low acoustic salience. The data also suggest that the relatively late stage of /ʂ/ acquisition by children may come about because of the weak acoustic information in the vowel for its distinction from /s/. 1 Introduction While there have been many studies in the last 30 years on the acoustic (Evers, Reetz & Lahiri 1998, Jongman, Wayland & Wong 2000,Nowak2006, Shadle 2006, Cheon & Anderson 2008, Maniwa, Jongman & Wade 2009), perceptual (McGuire 2007, Cheon & Anderson 2008,Li et al.
    [Show full text]
  • Detection and Classification of Whale Acoustic Signals
    Detection and Classification of Whale Acoustic Signals by Yin Xian Department of Electrical and Computer Engineering Duke University Date: Approved: Loren Nolte, Supervisor Douglas Nowacek (Co-supervisor) Robert Calderbank (Co-supervisor) Xiaobai Sun Ingrid Daubechies Galen Reeves Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Electrical and Computer Engineering in the Graduate School of Duke University 2016 Abstract Detection and Classification of Whale Acoustic Signals by Yin Xian Department of Electrical and Computer Engineering Duke University Date: Approved: Loren Nolte, Supervisor Douglas Nowacek (Co-supervisor) Robert Calderbank (Co-supervisor) Xiaobai Sun Ingrid Daubechies Galen Reeves An abstract of a dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Electrical and Computer Engineering in the Graduate School of Duke University 2016 Copyright c 2016 by Yin Xian All rights reserved except the rights granted by the Creative Commons Attribution-Noncommercial Licence Abstract This dissertation focuses on two vital challenges in relation to whale acoustic signals: detection and classification. In detection, we evaluated the influence of the uncertain ocean environment on the spectrogram-based detector, and derived the likelihood ratio of the proposed Short Time Fourier Transform detector. Experimental results showed that the proposed detector outperforms detectors based on the spectrogram. The proposed detector is more sensitive to environmental changes because it includes phase information. In classification, our focus is on finding a robust and sparse representation of whale vocalizations. Because whale vocalizations can be modeled as polynomial phase sig- nals, we can represent the whale calls by their polynomial phase coefficients.
    [Show full text]
  • Music Information Retrieval Using Social Tags and Audio
    A Survey of Audio-Based Music Classification and Annotation Zhouyu Fu, Guojun Lu, Kai Ming Ting, and Dengsheng Zhang IEEE Trans. on Multimedia, vol. 13, no. 2, April 2011 presenter : Yin-Tzu Lin (阿孜孜^.^) 2011/08 Types of Music Representation • Music Notation – Scores – Like text with formatting symbolic • Time-stamped events – E.g. Midi – Like unformatted text • Audio – E.g. CD, MP3 – Like speech 2 Image from: http://en.wikipedia.org/wiki/Graphic_notation Inspired by Prof. Shigeki Sagayama’s talk and Donald Byrd’s slide Intra-Song Info Retrieval Composition Arrangement probabilistic Music Theory Symbolic inverse problem Learning Score Transcription MIDI Conversion Accompaniment Melody Extraction Performer Structural Segmentation Synthesize Key Detection Chord Detection Rhythm Pattern Tempo/Beat Extraction Modified speed Onset Detection Modified timbre Modified pitch Separation Audio 3 Inspired by Prof. Shigeki Sagayama’s talk Inter-Song Info Retrieval • Generic-similar – Music Classification • Genre, Artist, Mood, Emotion… Music Database – Tag Classification(Music Annotation) – Recommendation • Specific-similar – Query by Singing/Humming – Cover Song Identification – Score Following 4 Classification Tasks • Genre Classification • Mood Classification • Artist Identification • Instrument Recognition • Music Annotation 5 Paper Outline • Audio Features – Low-level features – Middle-level features – Song-level feature representations • Classifiers Learning • Classification Task • Future Research Issues 6 Audio Features 7 Low-level Features
    [Show full text]
  • Selecting Proper Features and Classifiers for Accurate Identification of Musical Instruments
    International Journal of Machine Learning and Computing, Vol. 3, No. 2, April 2013 Selecting Proper Features and Classifiers for Accurate Identification of Musical Instruments D. M. Chandwadkar and M. S. Sutaone y The performance of the learning procedure is evaluated Abstract—Selection of effective feature set and proper by classifying new sound samples (cross-validation). classifier is a challenging task in problems where machine One of the most crucial aspects in the above procedure of learning techniques are used. In automatic identification of instrument classification is to find the right features [2] and musical instruments also it is very crucial to find the right set of classifiers. Most of the research on audio signal processing features and accurate classifier. In this paper, the role of various features with different classifiers on automatic has been focusing on speech recognition and speaker identification of musical instruments is discussed. Piano, identification. Few features used for these can be directly acoustic guitar, xylophone and violin are identified using applied to solve the instrument classification problem. various features and classifiers. Spectral features like spectral In this paper, feature extraction and selection for centroid, spectral slope, spectral spread, spectral kurtosis, instrument classification using machine learning techniques spectral skewness and spectral roll-off are used along with autocorrelation coefficients and Mel Frequency Cepstral is considered. Four instruments: piano, acoustic guitar, Coefficients (MFCC) for this purpose. The dependence of xylophone and violin are identified using various features instrument identification accuracy on these features is studied and classifiers. A number of spectral features, MFCCs, and for different classifiers. Decision trees, k nearest neighbour autocorrelation coefficients are used.
    [Show full text]
  • The Subjective Size of Melodic Intervals Over a Two-Octave Range
    JournalPsychonomic Bulletin & Review 2005, ??12 (?),(6), ???-???1068-1075 The subjective size of melodic intervals over a two-octave range FRANK A. RUSSO and WILLIAM FORDE THOMPSON University of Toronto, Mississauga, Ontario, Canada Musically trained and untrained participants provided magnitude estimates of the size of melodic intervals. Each interval was formed by a sequence of two pitches that differed by between 50 cents (one half of a semitone) and 2,400 cents (two octaves) and was presented in a high or a low pitch register and in an ascending or a descending direction. Estimates were larger for intervals in the high pitch register than for those in the low pitch register and for descending intervals than for ascending intervals. Ascending intervals were perceived as larger than descending intervals when presented in a high pitch register, but descending intervals were perceived as larger than ascending intervals when presented in a low pitch register. For intervals up to an octave in size, differentiation of intervals was greater for trained listeners than for untrained listeners. We discuss the implications for psychophysi- cal pitch scales and models of music perception. Sensitivity to relations between pitches, called relative scale describes the most basic model of pitch. Category pitch, is fundamental to music experience and is the basis labels associated with intervals assume a logarithmic re- of the concept of a musical interval. A musical interval lation between pitch height and frequency; that is, inter- is created when two tones are sounded simultaneously or vals spanning the same log frequency are associated with sequentially. They are experienced as large or small and as equivalent category labels, regardless of transposition.
    [Show full text]
  • Data-Driven Cepstral and Neural Learning of Features for Robust Micro-Doppler Classification
    Data-driven cepstral and neural learning of features for robust micro-Doppler classification Baris Erola, Mehmet Saygin Seyfioglub, Sevgi Zubeyde Gurbuzc, and Moeness Amina aCenter for Advanced Communication, Villanova University, Villanova, PA 19085, USA bDept. of Electrical-Electronics Eng., TOBB Univ. of Economics and Technology, Ankara, Turkey cDept. of Electrical and Computer Engineering, University of Alabama, AL 35487, USA ABSTRACT Automatic target recognition (ATR) using micro-Doppler analysis is a technique that has been a topic of great research over the past decade, with key applications to border control and security, perimeter defense, and force protection. Patterns in the movements of animals, humans, and drones can all be accomplished through classification of the target’s micro-Doppler signature. Typically, classification is based on a set of fixed, pre-defined features extracted from the signature; however, such features can perform poorly under low signal-to-noise ratio (SNR), or when the number and similarity of classes increases. This paper proposes a novel set of data-driven frequency-warped cepstral coefficients (FWCC) for classification of micro-Doppler signatures, and compares performance with that attained from the data-driven features learned in deep neural networks (DNNs). FWCC features are computed by first filtering the discrete Fourier Transform (DFT) of the input signal using a frequency-warped filter bank, and then computing the discrete cosine transform (DCT) of the logarithm. The filter bank is optimized for radar using genetic algorithms (GA) to adjust the spacing, weight, and width of individual filters. For a 11-class case of human activity recognition, it is shown that the proposed data-driven FWCC features yield similar classification accuracy to that of DNNs, and thus provides interesting insights on the benefits of learned features.
    [Show full text]
  • Psychoacoustics of Multichannel Audio
    THE PSYCHOACOUSTICS OF MULTICHANNEL AUDIO J. ROBERT STUART Meridian Audio Ltd Stonehill, Huntingdon, PE18 6ED England ABSTRACT This is a tutorial paper giving an introduction to the perception of multichannel sound reproduction. The important underlying psychoacoustic phenomena are reviewed – starting with the behaviour of the auditory periphery and moving on through binaural perception, central binaural phenomena and cognition. The author highlights the way the perception of a recording can be changed according to the number of replay channels used. The paper opens the question of relating perceptual and cognitive responses to directional sound or to sound fields. 1. INTRODUCTION 3.1 Pitch Multichannel systems are normally intended to present a The cochlea aids frequency selectivity by dispersing three-dimensional sound to the listener. In general, the excitation on the basilar membrane on a frequency- more loudspeakers that can be applied, the more accurately dependent basis. Exciting frequencies are mapped on a the sound field can be reproduced. Since all multichannel pitch scale roughly according to the dispersion (position) systems do not have a 1:1 relationship between transmitted and the integral of auditory filtering function. Several channels and loudspeaker feeds, a deep understanding of scales of pitch have been used; the most common being the human binaural system is necessary to avoid spatial, mel, bark and E.. Fig. 1 shows the inter-relationship loudness or timbral discrepancies. between these scales. This paper reviews some of the basic psychoacoustic The mel scale derives from subjective pitch testing and was mechanisms that are relevant to this topic. defined so that 1000mel º 1kHz.
    [Show full text]
  • Analyzing Noise Robustness of Mfcc and Gfcc Features in Speaker Identification
    ANALYZING NOISE ROBUSTNESS OF MFCC AND GFCC FEATURES IN SPEAKER IDENTIFICATION Xiaojia Zhao1and DeLiang Wang1,2 1Department of Computer Science and Engineering, The Ohio State University, USA 2Center for Cognitive Science, The Ohio State University, USA {zhaox, dwang}@cse.ohio-state.edu ABSTRACT prior to speaker recognition. The human ability to perform speaker recognition in Automatic speaker recognition can achieve a high level of noisy conditions has motivated studies of robust speaker performance in matched training and testing conditions. recognition from the perspective of computational auditory However, such performance drops significantly in mis- scene analysis. In one such study [4], we showed that a new matched noisy conditions. Recent research indicates that a speaker feature, gammatone frequency cepstral coefficients new speaker feature, gammatone frequency cepstral coeffi- (GFCC), shows superior noise robustness to commonly used cients (GFCC), exhibits superior noise robustness to com- mel-frequency cepstral coefficients (MFCC) in speaker iden- monly used mel-frequency cepstral coefficients (MFCC). To tification (SID) tasks (see also [5]). As for the reason, we gain a deep understanding of the intrinsic robustness of speculated that the front-end of GFCC, the gammatone GFCC relative to MFCC, we design speaker identification filterbank, might be more noise-robust than that of MFCC, experiments to systematically analyze their differences and the mel-filterbank. In particular, the frequency scales em- similarities. This study reveals that the nonlinear rectifica- ployed in the two filterbanks were believed to be the key tion accounts for the noise robustness differences primarily. difference although no convincing evidence was presented Moreover, this study suggests how to enhance MFCC ro- to support this hypothesis.
    [Show full text]
  • Lecture 6: Frequency Scales: Semitones, Mels, and Erbs
    Review Semitones Mel ERB Summary Lecture 6: Frequency Scales: Semitones, Mels, and ERBs Mark Hasegawa-Johnson ECE 417: Multimedia Signal Processing, Fall 2020 Review Semitones Mel ERB Summary 1 Review: Equivalent Rectangular Bandwidth 2 Musical Pitch: Semitones, and Constant-Q Filterbanks 3 Perceived Pitch: Mels, and Mel-Filterbank Coefficients 4 Masking: Equivalent Rectangular Bandwidth Scale 5 Summary Review Semitones Mel ERB Summary Outline 1 Review: Equivalent Rectangular Bandwidth 2 Musical Pitch: Semitones, and Constant-Q Filterbanks 3 Perceived Pitch: Mels, and Mel-Filterbank Coefficients 4 Masking: Equivalent Rectangular Bandwidth Scale 5 Summary Review Semitones Mel ERB Summary Fletcher's Model of Masking, Reviewed 1 The human ear pre-processes the audio using a bank of bandpass filters. 2 The power of the noise signal, in the filter centered at fc , is Z FS =2 2 Nfc = 2 R(f )jHfc (f )j df 0 3 2 The power of the tone is Tfc = A =2, if the tone is at frequency fc . 4 If there is any band in which Nfc + Tfc 10 log10 > 1dB Nfc then the tone is audible. Otherwise, not. Review Semitones Mel ERB Summary Equivalent rectangular bandwidth (ERB) The frequency resolution of your ear is better at low frequencies. In fact, the dependence is roughly linear (Glasberg and Moore, 1990): b ≈ 0:108f + 24:7 These are often called (approximately) constant-Q filters, because the quality factor is f Q = ≈ 9:26 b The dependence of b on f is not quite linear. A more precise formula is given in (Moore and Glasberg, 1983) as: f 2 f b = 6:23 + 93:39 + 28:52 1000 1000 Review Semitones Mel ERB Summary Equivalent rectangular bandwidth (ERB) By Dick Lyon, public domain image 2009, https://commons.wikimedia.org/wiki/File:ERB_vs_frequency.svg Review Semitones Mel ERB Summary The Gammatone Filter The shape of the filter is not actually rectangular.
    [Show full text]
  • Discrimination-Emphasized Mel-Frequency-Warping for Time-Varying Speaker Recognition
    APSIPA ASC 2011 Xi’an Discrimination-Emphasized Mel-Frequency-Warping for Time-Varying Speaker Recognition 1Linlin Wang, 2Thomas Fang Zheng, 3Chenhao Zhang and 4Gang Wang Center for Speech and Language Technologies, Division of Technical Innovation and Development, Tsinghua National Laboratory for Information Science and Technology Department of Computer Science and Technology, Tsinghua University, Beijing, 100084 E-mail: { 1wangll, 3zhangchh, 4wanggang }@cslt.riit.tsinghua.edu.cn, [email protected] Tel/Fax: +86-10-62796393 Abstract— Performance degradation with time varying is a voice just as that of fingerprints, they put forward this issue at generally acknowledged phenomenon in speaker recognition and the same time [3]. There was no evidence regarding the it is widely assumed that speaker models should be updated from stability of speaker-specific information throughout time. In time to time to maintain representativeness. However, it is costly, 1997, Sadaoki Furui summarized advances in automatic user-unfriendly, and sometimes, perhaps unrealistic, which speaker recognition in decades and also left the way to deal hinders the technology from practical applications. From a pattern recognition point of view, the time-varying issue in with long-term variability in people’s voice as an open speaker recognition requires such features that are speaker- question [1]. A similar idea was expressed in [4], where the specific, and as stable as possible across time-varying sessions. authors argued that a big challenge to uniquely characterize a Therefore, after searching and analyzing the most stable parts of person’s voice was that voice changes over time. feature space, a Discrimination-emphasized Mel-frequency- Performance degradation has also been observed in warping method is proposed.
    [Show full text]
  • Solution of the Three-Dimensional Inverse Acoustic Scattering Problem on the Basis of the Novikovðhenkin Algorithm N
    Acoustical Physics, Vol. 51, No. 4, 2005, pp. 367–375. Translated from AkusticheskiÏ Zhurnal, Vol. 51, No. 4, 2005, pp. 437–446. Original Russian Text Copyright © 2005 by Alekseenko, Burov, Rumyantseva. Solution of the Three-Dimensional Inverse Acoustic Scattering Problem on the Basis of the Novikov–Henkin Algorithm N. V. Alekseenko, V. A. Burov, and O. D. Rumyantseva Moscow State University, Vorob’evy gory, Moscow, 119992 Russia e-mail: [email protected] Received May 5, 2004 Abstract—A theoretical study of the practical abilities of the Novikov–Henkin algorithm, which is one of the most promising algorithms for solving three-dimensional inverse monochromatic scattering problems by func- tional analysis methods, is carried out. Numerical simulations are performed for model scatterers of different strengths in an approximation simplifying the reconstruction process. The resulting estimates obtained with the approximate algorithm prove to be acceptable for middle-strength scatterers. For stronger scatterers, an ade- quate reconstruction is possible on the basis of a rigorous solution. © 2005 Pleiades Publishing, Inc. METHODS FOR SOLVING INVERSE number of computing operations than in the iteration SCATTERING PROBLEMS methods. The functional analytical methods for solving one- The inverse scattering problem consists in the deter- dimensional inverse scattering problems appeared as mination of the characteristics of a scatterer located in ᑬ early as in the 1950s (Gel’fand, Levitan, Marchenko, a spatial domain and described by a function v(r), and others). The first comprehensive studies of the mul- ∈ ᑬ where r , from the measured scattered field. In the tidimensional inverse scattering problem were carried case of an infinite domain ᑬ, the function v(r) is out by Berezanskii in the 1950s and by Faddeev and assumed to be rapidly decreasing at infinity.
    [Show full text]