Label ground truth data for automated driving applications
The Ground Truth Labeler app enables you to label ground truth data in a video, in an image sequence, or from a custom data source reader. Using the app, you can:
Define rectangular regions of interest (ROI) labels, polyline ROI labels, pixel ROI labels, and scene labels, and use these labels to interactively label your ground truth data.
Use built-in detection or tracking algorithms to label your ground truth data.
Write, import, and use your own custom automation algorithm to automatically label ground truth. See Create Automation Algorithm for Labeling (Computer Vision System Toolbox).
Evaluate the performance of your label automation algorithms using a visual summary. See View Summary of Ground Truth Labels (Computer Vision System Toolbox).
Export the labeled ground truth as a
groundTruth object. You can use
this object for system verification or for training an object detector or
semantic segmentation network. See Train Object Detector or Semantic Segmentation Network from Ground Truth Data (Computer Vision System Toolbox).
Display time-synchronized signals, such as lidar or CAN bus data, using the
To learn more about the app, see Ground Truth Labeler App.
MATLAB® Toolstrip: On the Apps tab, under Automotive, click the app icon.
MATLAB command prompt: Enter
groundTruthLabeler opens a new session of the app, enabling
you to label ground truth data.
groundTruthLabeler(videoFileName) opens the app and loads the
input video. The video file must have an extension supported by
groundTruthLabeler(imageSeqFolder) opens the app and loads
the image sequence from the input folder.
imageSeqFolder must be
a string scalar or character vector that specifies the folder containing the image
groundTruthLabeler(imageSeqFolder,timestamps) opens the app
and loads a sequence of images with their corresponding timestamps.
timestamps must be a
duration vector of the same length as
the number of images in the sequence.
For example, load a sequence of road images and their corresponding timestamps into the app.
imageDir = fullfile(toolboxdir('driving'),'drivingdata','roadSequence'); load(fullfile(imageDir,'timeStamps.mat')) groundTruthLabeler(imageDir,timeStamps)
groundTruthLabeler(gtSource) opens the app and loads the
gtSource. The object contains a custom data source and
corresponding timestamps. See Use Custom Data Source Reader for Ground Truth Labeling (Computer Vision System Toolbox).
groundTruthLabeler(sessionFile) opens the app and loads a
saved app session,
sessionFile input contains the path and file name. The
sessionFile points to contains the saved
opens the app with a custom connector.
'connector' is a handle
The handle implements a custom analysis or visualization tool that is
time-synchronized with the Ground Truth Labeler app. For example, to
associate a connector target defined in class
For example, open the app, load a 10-second video into it, and open a lidar visualization tool that is time-synchronized to the video.
The built-in automation algorithms support the automation of rectangular ROI labels only. When you select a built-in algorithm and click Automate, scene labels, pixel labels, polyline labels, sublabels, and attributes are not imported into the automation session. To automate the labeling of these features, create a custom automation algorithm. See Create Automation Algorithm for Labeling (Computer Vision System Toolbox).
Pixel ROI labels do not support sublabels or attributes.
The Label Summary window does not support sublabels or attributes
To avoid having to relabel ground truth with new labels, organize the labeling scheme you want to use before marking your ground truth.
The Ground Truth Labeler app provides built-in algorithms that you can use to automate labeling. From the app toolstrip, click Select Algorithm, and then select an automation algorithm.
|Built-In Automation Algorithm||Description|
ACF People Detector
|Detect and label people using a pretrained detector based on aggregate channel features (ACF). With this algorithm, you do not need to draw any ROI labels.|
|Track and label one or more rectangular ROI labels over short intervals using the Kanade-Lucas-Tomasi (KLT) algorithm.|
|Estimate ROIs in intermediate frames using the interpolation of rectangular ROIs in key frames. Draw ROIs on a minimum of two frames (at the beginning and at the end of the interval). The interpolation algorithm estimates the ROIs between the frames.|
ACF Vehicle Detector
|Detect and label vehicles using a pretrained detector based on ACF. With this algorithm, you do not need to draw any ROI labels.|