This course concerns with analysis and processing of speech signals for development of different voice biometric based applications. The course will provide hands-on experience to the participants about various tasks involved in the analysis of speech signal for extraction of different information, detection of different events in speech signal, development of speech enhancement systems, development of speech recognition systems, speaker recognition systems, speaker diarization systems, language identification systems and their applications for different real word applications. The course will also include several invited talks from the leading experts working in different application areas of speech and Language processing. Such invited talks may ignite the research community to look the speech and Language processing technology from a different perspective.Course Outcome:
Upon successful completion of this course, students, faculties and researchers should be able to understand the following:
Speech production and perception, Information sources in speech, Linguistic aspect of speech, Acoustic and articulatory phonetics, Nature of speech signal, Models for speech analysis.
Overview of Fourier representation, Short-term Fourier transform (STFT), Filter-bank views of STFT, Time, Frequency and Time-Frequency analysis.
Basis and Development, Homomorphic signal processing, Real and Complex cepstrum, Mel-frequency cepstral coefficient (MFCC), Delta and Delta-Delta.
Basis and Development, Levinson-Durbin’s method, Normalized error, LP spectrum, LP cepstrum, LP residual.
Dynamic Time Warping (DTW), Vector Quantization (VQ), Gaussian Mixture Model (GMM), GMM-Universal Background Model (UBM), Hidden Markov Model (HMM), N-grams, Artificial Neural Network (ANN), Support Vector Machine (SVM), Joint Factor Analysis, I-vector.
Objective, Issues, Development of speech enhancement system by spectral, temporal processing methods.
Objective, Issues, Block diagram description, Classification, Development of text-dependent, text-independent and voice password based speaker identification and verification systems.
Objective, Issues, Block diagram description, Development of speaker diarization systems.
Objective, Issues, Block diagram description, Development of speech recognition systems.
Objective, Issues, Block diagram description, development of Language identification systems.