Main Content

Preprocess Data for Deep Neural Networks

Manage and preprocess data for deep learning

Preprocessing data to ensure that it is in a format that the network can accept is a common first step in deep learning workflows. For example, you can resize image input to match the size of an image input layer. You can also preprocess data to enhance desired features or reduce artifacts that can bias the network. For example, you can normalize or remove noise from input data.

You can preprocess image input with operations such as resizing by using datastores and functions available in MATLAB® and Deep Learning Toolbox™. Other MATLAB toolboxes offer functions, datastores, and apps for labeling, processing, and augmenting deep learning data. Use specialized tools from other MATLAB toolboxes to process data for domains such as image processing, object detection, semantic segmentation, signal processing, audio processing, and text analytics.

Apps

Image LabelerLabel images for computer vision applications
Video LabelerLabel video for computer vision applications
Ground Truth LabelerLabel ground truth data for automated driving applications
Lidar LabelerLabel ground truth data in lidar point clouds (Seit R2020b)
Signal LabelerLabel signal attributes, regions, and points of interest, and extract features (Seit R2019a)

Funktionen

imageDatastoreDatastore for image data
augmentedImageDatastoreTransform batches to augment image data
imageDataAugmenterConfigure image data augmentation
augmentApply identical random transformations to multiple images
minibatchqueueCreate mini-batches for deep learning (Seit R2020b)

Themen

Preprocess Deep Learning Data

Customize Datastores

Label Ground Truth Training Data