Main Content

Work with Ground Truth Data

Select, merge, and load training data for deep learning

Select labels, change file paths to the data, and merge ground truth objects. You can share ground truth objects with labeling colleagues to continue labeling and training.

To share and review image labeling with colleagues, consider creating a project within the Image Labeling app. For more details, see Get Started with Team-Based Labeling.


Image LabelerLabel images for computer vision applications
Video LabelerLabel video for computer vision applications


expand all

selectLabelsByGroupSelect ground truth labels by label group
selectLabelsByTypeSelect ground truth labels by label type
selectLabelsByNameSelect ground truth labels by label name
groundTruthGround truth label data
mergeMerge two or more ground truth objects (Since R2023b)
pixelLabelDatastoreDatastore for pixel label data
boxLabelDatastoreDatastore for bounding box label data (Since R2019b)
changeFilePathsChange file paths in ground truth data
writeVideoScenesWrite video sequence to video file (Since R2021b)
sceneTimeRangesTime ranges of scene labels from ground truth data (Since R2021b)
gatherLabelDataGather label data from ground truth (Since R2021a)
objectDetectorTrainingDataCreate training data for an object detector
pixelLabelTrainingDataCreate training data for semantic segmentation from ground truth
polyToBlockedImageCreate labeled blockedImage object from set of ROIs (Since R2021b)
mergeMerge two or more ground truth objects (Since R2023b)
attributeTypeAttribute type enumerations for labeling
labelTypeLabel type enumerations for labeling