Course Content
-
Day 01 - Module 1 Computer Speech Processing 0/3
-
Overview of the courseLecture1.1
-
A Brief history of speech processingLecture1.2
-
The progress in three decades of research.Lecture1.3
-
-
Module-2 Speech Sounds 0/3
-
Physiology of speech production;Lecture2.1
-
A model for speech productionLecture2.2
-
Types of soundsLecture2.3
-
-
Module-3 Speech Processing 0/3
-
Feature extractionLecture3.1
-
Linear predictive coding;Lecture3.2
-
Cepstral coefficients.Lecture3.3
-
-
Module-4 Classification & Machine Learning 0/3
-
Classification of static vectorsLecture4.1
-
Popular classifiers in machine learning;Lecture4.2
-
Bayes decision ruleLecture4.3
-
-
Day 02 - Module-1 Multivariate Distributions 0/3
-
Distributions for modeling acoustics;Lecture5.1
-
ML parameter estimationLecture5.2
-
Gaussian mixture models.Lecture5.3
-
-
Module-2 Elementary Pattern Recognition 0/4
-
Template matchingLecture6.1
-
Dynamic programmingLecture6.2
-
Dynamic time warpingLecture6.3
-
Morphing dynamic time warping to hidden Markov models.Lecture6.4
-
-
Module-3 HMM Parameter Estimation 0/2
-
EM algorithm for HMMLecture7.1
-
Viterbi algorithm.Lecture7.2
-
-
Module- 5 Finite State Machines 0/3
-
Day 03 - Module-1 Basic ASR Systems 0/5
-
Increasing the complexity in stepsLecture9.1
-
Context-independent to context-dependent acoustic models;Lecture9.2
-
Gaussian mixture splittingLecture9.3
-
Multistage searchLecture9.4
-
Simple methods for system combinationLecture9.5
-
-
Module-2 Distribution Clustering 0/3
-
Need for context-dependent models;Lecture10.1
-
Clustering distributionsLecture10.2
-
Decision tree based state clustering; from decision tree to transducers.Lecture10.3
-
-
Module-3 Discriminative Training 0/3
-
Module-4 Speaker Adaptation 0/6
-
Bridging the gap between SI and SD systemsLecture12.1
-
Types of adaptation – speaker adaptive training and speaker adaptationLecture12.2
-
MLLRLecture12.3
-
constrained MLLRLecture12.4
-
Regression classLecture12.5
-
MAP.Lecture12.6
-
-
Day 04 - Module-1 Feature Space Normalization 0/3
-
Vocal tract length normalizationLecture13.1
-
All pass transformLecture13.2
-
Maximum likelihood linear transform under different constraints.Lecture13.3
-
-
Module-2 Large Vocabulary ASR 0/3
-
Sketch of a state-of-the-art LV-ASRLecture14.1
-
Cross adaptationLecture14.2
-
System combination.Lecture14.3
-
-
Module -3 Advanced Topics 0/4
-
Overview of robust ASR techniquesLecture15.1
-
Non-parametric techniquesLecture15.2
-
Utterance classificationLecture15.3
-
Detecing para-/extra-linguistic features in speech.Lecture15.4
-
This content is protected, please login and enroll course to view this content!
Prev
Dynamic time warping
Next
EM algorithm for HMM