A5027129423- Structural Health Monitoring Techniques
- Neural Networks and Applications
- Speech and Audio Processing
- EEG and Brain-Computer Interfaces
- Non-Invasive Vital Sign Monitoring
- ECG Monitoring and Analysis
- Blind Source Separation Techniques
- Heart Rate Variability and Autonomic Control
- Advanced Adaptive Filtering Techniques
Hanyang University
2021
In this paper, we propose a novel data augmentation technique employing multivariate Gaussian distribution (DA-MGD) for neural network (NN)-based blood pressure (BP) estimation, which incorporates the relationship between features in multi-dimensional feature vector to describe correlated real-valued random variables successfully. To verify proposed algorithm against conventional algorithm, compare results terms of mean error (ME) with standard deviation and Pearson correlation using 110...
This study develops the deep Q-network (DQN)-based noise suppression for robust speech recognition purposes under ambient noise. We thus design a reinforcement algorithm that combines DQN training with neural networks (DNN) to let learning (RL) work complex and high dimensional environments like recognition. For this, we elaborate on choose best action is quantized gain by observation of noisy signal rewards including both word error rate (WER) objective quality measure. Experiments...