CNN+LSTM Architecture for Speech Emotion Recognition with Data Augmentation
arXiv:1802.05630 · doi:10.21437/SMM.2018-5
Abstract
In this work we design a neural network for recognizing emotions in speech, using the IEMOCAP dataset. Following the latest advances in audio analysis, we use an architecture involving both convolutional layers, for extracting high-level features from raw spectrograms, and recurrent ones for aggregating long-term dependencies. We examine the techniques of data augmentation with vocal track length perturbation, layer-wise optimizer adjustment, batch normalization of recurrent layers and obtain highly competitive results of 64.5% for weighted accuracy and 61.7% for unweighted accuracy on four emotions.
arXiv comment5 pages, 3 figures
References in corpus (5)
- Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
- Convolutional LSTM Network: A Machine Learning Approach for Precipitation Nowcasting
- Deep Speech 2: End-to-End Speech Recognition in English and Mandarin
- Layer Normalization
- Gradient Normalization & Depth Based Decay For Deep Learning
Cited by in corpus (10)
- Speech Emotion Recognition with Dual-Sequence LSTM Architecture
- Exploring Wav2vec 2.0 fine-tuning for improved speech emotion recognition
- Emotion Recognition In Persian Speech Using Deep Neural Networks
- A novel policy for pre-trained Deep Reinforcement Learning for Speech Emotion Recognition
- A Systematic Evaluation of Adversarial Attacks against Speech Emotion Recognition Models
- Multi-Window Data Augmentation Approach for Speech Emotion Recognition
- Enhancing Generalization in PPG-Based Emotion Measurement with a CNN-TCN-LSTM Model
- Enhancing Segment-Based Speech Emotion Recognition by Deep Self-Learning
- Speech Emotion Recognition Using Deep Sparse Auto-Encoder Extreme Learning Machine with a New Weighting Scheme and Spectro-Temporal Features Along with Classical Feature Selection and A New Quantum-Inspired Dimension Reduction Method
- Speech Emotion Recognition System by Quaternion Nonlinear Echo State Network