ICASSP99 Speaker Recognition

Speaker Recognition
Home Full List of Titles 1: Speech Processing CELP Coding Large Vocabulary Recognition Speech Analysis and Enhancement Acoustic Modeling I ASR Systems and Applications Topics in Speech Coding Speech Analysis Low Bit Rate Speech Coding I Robust Speech Recognition in Noisy Environments Speaker Recognition Acoustic Modeling II Speech Production and Synthesis Feature Extraction Robust Speech Recognition and Adaptation Low Bit Rate Speech Coding II Speech Understanding Language Modeling I 2: Speech Processing, Audio and Electroacoustics, and Neural Networks Acoustic Modeling III Lexical Issues/Search Speech Understanding and Systems Speech Analysis and Quantization Utterance Verification/Acoustic Modeling Language Modeling II Adaptation /Normalization Speech Enhancement Topics in Speaker and Language Recognition Echo Cancellation and Noise Control Coding Auditory Modeling, Hearing Aids and Applications of Signal Processing to Audio and Acoustics Spatial Audio Music Applications Application - Pattern Recognition & Speech Processing Theory & Neural Architecture Signal Separation Application - Image & Nonlinear Signal Processing 3: Signal Processing Theory & Methods I Filter Design and Structures Detection Wavelets Adaptive Filtering: Applications and Implementation Nonlinear Signals and Systems Time/Frequency and Time/Scale Analysis Signal Modeling and Representation Filterbank and Wavelet Applications Source and Signal Separation Filterbanks Emerging Applications and Fast Algorithms Frequency and Phase Estimation Spectral Analysis and Higher Order Statistics Signal Reconstruction Adaptive Filter Analysis Transforms and Statistical Estimation Markov and Bayesian Estimation and Classification 4: Signal Processing Theory & Methods II, Design and Implementation of Signal Processing Systems, Special Sessions, and Industry Technology Tracks System Identification, Equalization, and Noise Suppression Parameter Estimation Adaptive Filters: Algorithms and Performance DSP Development Tools VLSI Building Blocks DSP Architectures DSP System Design Education Recent Advances in Sampling Theory and Applications Steganography: Information Embedding, Digital Watermarking, and Data Hiding Speech Under Stress Physics-Based Signal Processing DSP Chips, Architectures and Implementations DSP Tools and Rapid Prototyping Communication Technologies Image and Video Technologies Automotive Applications / Industrial Signal Processing Speech and Audio Technologies Defense and Security Applications Biomedical Applications Voice and Media Processing Adaptive Interference Cancellation 5: Communications, Sensor Array and Multichannel Source Coding and Compression Compression and Modulation Channel Estimation and Equalization Blind Multiuser Communications Signal Processing for Communications I CDMA and Space-Time Processing Time-Varying Channels and Self-Recovering Receivers Signal Processing for Communications II Blind CDMA and Multi-Channel Equalization Multicarrier Communications Detection, Classification, Localization, and Tracking Radar and Sonar Signal Processing Array Processing: Direction Finding Array Processing Applications I Blind Identification, Separation, and Equalization Antenna Arrays for Communications Array Processing Applications II 6: Multimedia Signal Processing, Image and Multidimensional Signal Processing, Digital Signal Processing Education Multimedia Analysis and Retrieval Audio and Video Processing for Multimedia Applications Advanced Techniques in Multimedia Video Compression and Processing Image Coding Transform Techniques Restoration and Estimation Image Analysis Object Identification and Tracking Motion Estimation Medical Imaging Image and Multidimensional Signal Processing Applications I Segmentation Image and Multidimensional Signal Processing Applications II Facial Recognition and Analysis Digital Signal Processing Education Author Index A B C D E F G H I J K L M N O P Q R S T U V W X Y Z	Speaker Verification Performance And The Length Of Test Sentence Authors: Jialong He, Li Liu, Page (NA) Paper number 1021 Abstract: It is known that the performance of a speaker verification system improves with the length of test sentences. However, little is known about the exact relation between the performance and the test length. That makes it difficult to compare the results from various studies in which different test lengths have been used to evaluate the systems. In this paper, we have proposed a method to calculate the verification error rates at any lengths of test sentences, as long as the error rates at two different lengths are given. The accuracy of this calculation method is demonstrated with a speaker verification experiment and with the results reported in literature. Good agreement is shown between the calculated values and that measured through experiments. IC991021.PDF (From Author) IC991021.PDF (Rasterized) TOP On The Use Of Some Divergence Measures In Speaker Recognition Authors: Rivarol Vergin, Douglas O'Shaughnessy, Page (NA) Paper number 1336 Abstract: The first motivation for using Gaussian Mixture Models for text-independent speaker identification is based on the observation that a linear combination of Gaussian basis functions is capable of representing a large class of sample distributions. While this technique gives generally good results little is known about which specific part of a speech signal bests identifies a speaker. This contribution suggests a procedure, based on Jensen divergence measure, to automatically extract from the input speech signal the part that best contributes to identify a speaker. Experiments conducted using the Spidre database indicate a significant improvement in the performance of the speaker recognition system. IC991336.PDF (From Author) IC991336.PDF (Rasterized) TOP Improving a GMM Speaker Verification System by Phonetic Weighting Authors: Roland Auckenthaler, Eluned S Parris, Michael J Carey, Page (NA) Paper number 1440 Abstract: This paper compares two approaches to speaker verification, Gaussian mixture models (GMMs) and Hidden Markov models (HMMs). The GMM based system outperformed the HMM system, this was mainly due to the ability of the GMM to make better use of the training data. The best scoring GMM frames were strongly correlated with particular phonemes e.g. vowels and nasals. Two techniques were used to try and exploit the different amounts of discrimination provided by the phonemes to improve the performance of the GMM based system. Applying linear weighting to the phonemes showed that less than half of the phonemes were contributing to the overall system performance. Using an MLP to weight the phonemes provided a significant improvement in performance for male speakers but no improvement has yet been achieved for women. IC991440.PDF (From Author) IC991440.PDF (Rasterized) TOP A Hybrid Score Measurement For HMM-Based Speaker Verification Authors: Yong Gu, Vocalis Ltd., UK (U.K.) Trevor Thomas, Vocalis Ltd., UK (U.K.) Page (NA) Paper number 1636 Abstract: In speaker verification the world model based approach and the cohort model based approach have been used for better HMM score measurements for verification comparison. From theoretical analysis these two approaches represent two different paradigms for verification decision-making strategy. Two techniques could be combined for a better solution. In the paper we present a hybrid score measurement which combines the world model based technique and the cohort model based technique together. The method is evaluated with the YOHO database. The results show that the combination can lead a better score measurement which improves speaker verification performance. An experimental comparison between the world model based approach and the cohort model based approach with the YOHO database can also be found in the paper. IC991636.PDF (From Author) IC991636.PDF (Rasterized) TOP Polynomial Classifier Techniques for Speaker Verification Authors: William M Campbell, Khaled T Assaleh, Page (NA) Paper number 1735 Abstract: Modern speaker verification applications require high accuracy at low complexity. We propose the use of a polynomial-based classifier to achieve this objective. We demonstrate a new combination of techniques which makes polynomial classification accurate and powerful for speaker verification. We show that discriminative training of polynomial classifiers can be performed on large data sets. A prior probability compensation method is detailed which increases accuracy and normalizes the output score range. Results are given for the application of the new methods to YOHO. IC991735.PDF (From Author) IC991735.PDF (Rasterized) TOP Channel-Robust Speaker Identification using Modified-Mean Cepstral Mean Normalization with Frequency Warping Authors: Alvin A Garcia, Richard J Mammone, Page (NA) Paper number 2165 Abstract: The performance of automatic speaker recognition systems is significantly degraded by acoustic mismatches between training and testing conditions. Such acoustic mismatches are commonly encountered in systems that operate on speech collected over telephone networks, where different handsets and different network routes impose varying convolutional distortions on the speech signal. A new algorithm, the Modified-Mean Cepstral Mean Normalization with Frequency Warping (MMCMNFW) method, which improves upon the commonly-employed Cepstral Mean Subtraction method, has been developed. Experimental results on closed-set speaker identification tasks on a channel-corrupted subset of the TIMIT database and on a subset of the NTIMIT database are presented. The new algorithm is shown to offer improved recognition rates over other existing channel normalization methods on these databases. IC992165.PDF (From Author) IC992165.PDF (Rasterized) TOP Feature Selection Using Genetics-Based Algorithm and Its Application to Speaker Identification Authors: Mubeccel Demirekler, Ali Haydar, Page (NA) Paper number 5026 Abstract: This paper introduces the use of genetics-based algorithm in the reduction of 24 parameter set (i.e. the base set) to a 5, 6, 7, 8 or 10 parameter set, for each speaker in text-independent speaker identification. The feature selection is done by finding the best features that discriminates a person from his/her two closest neighbors. The experimental results show that there is approximately 5% increase in the recognition rate when the reduced set of parameters are used. Also the amount of calculation necessary for speaker recognition using the reduced set of features is much less than the amount of calculation required using the complete feature set in the testing phase. Hence it is more desirable to use the subset of the complete feature set found using the genetic algorithm suggested. IC995026.PDF (Scanned) TOP