Multiple speaker localization using mixture of Gaussian model with manifold-based centroids

Keyphrases

• Relative Transfer Function 100%
• Centroid 100%
• Speaker Localization 100%
• Mixture of Gaussian Model 100%
• Utterance 33%
• Small Sets 16%
• Sparsity 16%
• Feature Vector 16%
• Large Set 16%
• Speech Signal 16%
• Reverberant Enclosure 16%
• Reverberation 16%
• Expectation Maximization 16%
• Maximum Likelihood Estimation 16%
• Inference Methods 16%
• Training Stages 16%
• Single-speaker 16%
• Localization Accuracy 16%
• Intermittency 16%
• Semi-supervised Learning 16%
• Multiple Speakers 16%
• Propagation Model 16%
• Source Location 16%
• Source Position 16%
• Multiple Manifolds 16%
• Data-driven Approach 16%
• Function Location 16%
• Nonlinear Manifold 16%
• Localized Source 16%
• Microphone Position 16%
• Simple Propagation 16%

Computer Science

• Transfer Function 100%
• Gaussian Model 100%
• Sparsity 16%
• maximum-likelihood 16%
• Baseline Method 16%
• Semisupervised Learning 16%
• Driven Approach 16%
• Feature Vector 16%
• Propagation Model 16%
• Source Location 16%
• Likelihood Estimation 16%