Classify heartbeat electrocardiogram data using deep learning and the continuous wavelet transform.

Classify human electrocardiogram (ECG) signals using wavelet time scattering and a support vector machine (SVM) classifier. In wavelet scattering, data is propagated through a series of wavelet transforms, nonlinearities, and averaging to produce low-variance representations of time series. Wavelet time scattering yields signal representations insensitive to shifts in the input signal without sacrificing class discriminability. You must have the Wavelet Toolbox™ and the Statistics and Machine Learning Toolbox™ to run this example. The data used in this example are publicly available from PhysioNet. You can find a deep learning approach to this classification problem in this example Classify Time Series Using Wavelet Analysis and Deep Learning and a machine learning approach in this example Signal Classification Using Wavelet-Based Features and Support Vector Machines.

Classify the genre of a musical excerpt using wavelet time scattering and the audio datastore. In wavelet scattering, data is propagated through a series of wavelet transforms, nonlinearities, and averaging to produce low-variance representations of the data. These low-variance representations are then used as inputs to a classifier.

Accelerate the computation of wavelet scattering features using gpuArray and a parallelized "depth-first" version of wavelet time scattering. You must have a CUDA-enabled NVIDIA GPU with compute capability 3.0 or higher. See GPU Support by Release for details.

Classify human phonocardiogram (PCG) recordings using wavelet time scattering and a support vector machine (SVM) classifier. Phonocardiograms are acoustic recordings of sounds produced by the systolic and diastolic phases of the heart. Auscultation of the heart continues to play an important diagnostic role in assessing cardiac health. Unfortunately, many areas of the world lack sufficient numbers of medical personnel trained in heart auscultation. Accordingly, it is necessary to develop reliable automated ways of interpreting phonocardiogram data.

Classify spoken digits using both machine and deep learning techniques. In the example, you perform classification using wavelet time scattering with a support vector machine (SVM) and with a long short-term memory (LSTM) network. You also apply Bayesian optimization to determine suitable hyperparameters to improve the accuracy of the LSTM network. In addition, the example illustrates an approach using a deep convolutional neural network (CNN) and mel-frequency spectrograms.

Accelerate the computation of wavelet scattering features using gpuArray and a parallelized "depth-first" version of wavelet time scattering. You must have a CUDA-enabled NVIDIA GPU with compute capability 3.0 or higher. See GPU Support by Release for details. This example uses an NVIDIA Titan XP GPU with compute capability 6.1.

Create a multi-model late fusion system for acoustic scene recognition. The example trains a convolutional neural network (CNN) using mel spectrograms and an ensemble classifier using wavelet scattering. The example uses the TUT dataset for training and evaluation [1].

Classify human electrocardiogram (ECG) signals using wavelet-based feature extraction and a support vector machine (SVM) classifier. The problem of signal classification is simplified by transforming the raw ECG signals into a much smaller set of features that serve in aggregate to differentiate different classes. You must have Wavelet Toolbox™, Signal Processing Toolbox™, and Statistics and Machine Learning Toolbox™ to run this example. The data used in this example are publicly available from PhysioNet.

Classify textures using wavelet image scattering. In addition to Wavelet Toolbox™, this example also requires Parallel Computing Toolbox™ and Image Processing Toolbox™.

Use wavelet scattering for image classification. This example requires Wavelet Toolbox™, Deep Learning Toolbox™, and Parallel Computing Toolbox™.

Classify radar returns with both machine and deep learning approaches. The machine learning approach uses wavelet scattering feature extraction coupled with a support vector machine. Additionally, two deep learning approaches are illustrated: transfer learning using SqueezeNet and a Long Short-Term Memory (LSTM) recurrent neural network. Note that the data set used in this example does not require advanced techniques but the workflow is described because the techniques can be extended to more complex problems.

Classify radar waveform types of generated synthetic data using the Wigner-Ville distribution (WVD) and a deep convolutional neural network (CNN).

Accelerate scalogram computation using GPUs. The computed scalograms are used as the input features to deep convolution neural networks (CNN) for ECG and spoken digit classification.

The workflow to classify human electrocardiogram (ECG) signals using the Continuous Wavelet Transform (CWT) and a deep convolutional neural network (CNN). This example also provides information on how to generate and deploy the code and CNN for prediction on a Raspberry Pi target (ARM®-based device).

Use wavelet and deep learning techniques to detect transverse pavement cracks and localize their position. The example demonstrates the use of wavelet scattering sequences as inputs to a gated recurrent unit (GRU) and 1-D convolutional network to classify time series based on the presence or absence of a crack. The data are vertical acceleration measurements obtained from a sensor mounted on the suspension knuckle of the front passenger seat wheel. Early identification of developing transverse cracks is important for pavement performance evaluation and maintenance. Reliable automatic detection methods enable more frequent and extensive monitoring.

Generate and deploy a CUDA® executable that classifies human electrocardiogram (ECG) signals using features extracted by the continuous wavelet transform (CWT) and a pretrained convolutional neural network (CNN).