AI for Signals

Signal labeling, feature engineering, dataset generation, anomaly detection

Signal Processing Toolbox™ provides functionality to perform signal labeling, feature engineering, and dataset generation for machine learning and deep learning workflows. The toolbox also offers an autoencoder object that you can train and use to detect anomalies in signal data.

Apps

Signal Analyzer	Visualize and compare multiple signals and spectra
Signal Labeler	Label signal attributes, regions, and points of interest, and extract features
EDF File Analyzer	View EDF or EDF+ files (Since R2021a)
Experiment Manager	Design and run experiments to train and compare deep learning networks (Since R2020a)

Functions

expand all

Signal Labeling

`labeledSignalSet`	Create labeled signal set
`signalLabelDefinition`	Create signal label definition
`countlabels`	Count number of unique labels (Since R2021a)
`filenames2labels`	Get list of labels from filenames (Since R2022b)
`folders2labels`	Get list of labels from folder names (Since R2021a)
`framelbl`	Partition label sequence into frames (Since R2024a)
`framesig`	Partition signal into frames (Since R2024a)
`splitlabels`	Find indices to split labels according to specified proportions (Since R2021a)
`signalMask`	Modify and convert signal masks and extract signal regions of interest (Since R2020b)
`binmask2sigroi`	Convert binary mask to matrix of ROI limits (Since R2020b)
`extendsigroi`	Extend signal regions of interest to left and right (Since R2020b)
`extractsigroi`	Extract signal regions of interest (Since R2020b)
`mergesigroi`	Merge signal regions of interest (Since R2020b)
`removesigroi`	Remove signal regions of interest (Since R2020b)
`shortensigroi`	Shorten signal regions of interest from left and right (Since R2020b)
`sigroi2binmask`	Convert matrix of ROI limits to binary mask (Since R2020b)
`sigrangebinmask`	Label signal samples with values within a specified range (Since R2023a)

Datastores and Data Management

`edfinfo`	Get information about EDF/EDF+ file (Since R2020b)
`edfwrite`	Create or modify EDF or EDF+ file (Since R2021a)
`edfheader`	Create header structure for EDF or EDF+ file (Since R2021a)
`edfread`	Read data from EDF/EDF+ file (Since R2020b)
`signalDatastore`	Datastore for collection of signals (Since R2020a)
`resize`	Resize data by adding or removing elements (Since R2023b)
`paddata`	Pad data by adding elements (Since R2023b)
`trimdata`	Trim data by removing elements (Since R2023b)

Feature Extraction

`findchangepts`	Find abrupt changes in signal
`findpeaks`	Find local maxima
`tfridge`	Time-frequency ridges
`instbw`	Estimate instantaneous bandwidth (Since R2021a)
`instfreq`	Estimate instantaneous frequency
`powerbw`	Power bandwidth
`pspectrum`	Analyze signals in the frequency and time-frequency domains
`spectralCrest`	Spectral crest for signals and spectrograms
`spectralEntropy`	Spectral entropy for signals and spectrograms
`spectralFlatness`	Spectral flatness for signals and spectrograms
`spectralKurtosis`	Spectral kurtosis for signals and spectrograms
`spectralSkewness`	Spectral skewness for signals and spectrograms
`scalarFeatureOptions`	Store information for converting feature vectors to scalar values (Since R2024a)
`signalFrequencyFeatureExtractor`	Streamline signal frequency feature extraction (Since R2021b)
`frequencyScalarFeatureOptions`	Store information for converting frequency-domain feature vectors to scalar values (Since R2024b)
`signalTimeFeatureExtractor`	Streamline signal time feature extraction (Since R2021a)
`timeScalarFeatureOptions`	Store information for converting time-domain feature vectors to scalar values (Since R2024b)
`signalTimeFrequencyFeatureExtractor`	Streamline signal time-frequency feature extraction (Since R2024a)
`timeFrequencyScalarFeatureOptions`	Store information for converting time-frequency-domain feature vectors to scalar values (Since R2024a)
`zerocrossrate`	Zero-crossing rate (Since R2021b)

Differentiable Signal Processing

`dlstft`	Deep learning short-time Fourier transform (Since R2021a)
`dlistft`	Deep learning inverse short-time Fourier transform (Since R2024a)
`stftLayer`	Short-time Fourier transform layer (Since R2021b)
`istftLayer`	Inverse short-time Fourier transform layer (Since R2024a)

Anomaly Detection

deepSignalAnomalyDetector Create signal anomaly detector (Since R2023a)

Topics

Accelerate Signal Feature Extraction and Classification Using a Parallel Pool of Workers
Use parallel computing to extract signal multidomain features for bearing fault detection.
Manage Data Sets for Machine Learning and Deep Learning Workflows
Organize, access, and manage data sets for different AI applications.
Choose an App to Label Ground Truth Data
Decide which app to use to label ground truth data: Image Labeler, Video Labeler, Ground Truth Labeler, Lidar Labeler, Signal Labeler, or Medical Image Labeler.
Radar and Communications Waveform Classification Using Deep Learning (Phased Array System Toolbox)
Classify radar and communications waveforms using the Wigner-Ville distribution (WVD) and a deep convolutional neural network (CNN).
Label Radar Signals with Signal Labeler (Radar Toolbox)
Label the time and frequency features of pulse radar signals with added noise. (Since R2021a)
Label Signal Attributes, Regions of Interest, and Points
Use Signal Labeler to label attributes, regions, and points of interest in a set of whale songs.
Automate Signal Labeling with Custom Functions
Use Signal Labeler to locate and label QRS complexes and R peaks of ECG signals.
Generate Synthetic Signals Using Conditional GAN
Use a conditional generative adversarial network to produce synthetic signals.
Waveform Segmentation Using Deep Learning
Segment human electrocardiogram signals using time-frequency analysis and deep learning.
Pedestrian and Bicyclist Classification Using Deep Learning (Radar Toolbox)
Classify pedestrians and bicyclists based on their micro-Doppler characteristics using deep learning and time-frequency analysis. (Since R2021a)
Wavelet Time Scattering Classification of Phonocardiogram Data (Wavelet Toolbox)
Classify human phonocardiogram recordings using wavelet time scattering and a support vector machine classifier.
Anomaly Detection Using Autoencoder and Wavelets
Use wavelet-extracted features and an autoencoder to detect arc signals in a DC system.
Detect Anomalies in ECG Data Using Wavelet Scattering and LSTM Autoencoder in Simulink (DSP System Toolbox)
Use wavelet scattering and deep learning network to detect anomalies in ECG signals.
Train Spoken Digit Recognition Network Using Out-of-Memory Features
Train a spoken digit recognition network on out-of-memory auditory spectrograms using a transformed datastore.
Denoise Speech Using Deep Learning Networks
Denoise speech signals using fully connected and convolutional neural networks.
Classify Time Series Using Wavelet Analysis and Deep Learning
Classify ECG signals using the continuous wavelet transform and a deep convolutional neural network.
Spectral Descriptors (Audio Toolbox)
Overview and applications of spectral descriptors.