Combining Multiple Ground Truths

I am trying to do multiple object detection using trainXXXObjectDetector family of functions. Problem is ground truth comes from a set of video files. Video labelling app only allows a single video to be marked in each session. What is the correct procedure for combining multiple ground truth objects in to a combined ground truth object. Say I have one ground truth for objectA marked using videoA and another ground truth containing objectB with videoB how do i create a ground truth object that contains combined data for both so i can train a trainXXXObjectDetector for both objectA and objectB

Answer 1

[Birju Patel]

You don't need to combine the groundTruth objects. Use objectDetectorTrainingData to extract training data from multiple groundTruth objects. You can pass in an array of groundTruth objects into this function.

See https://www.mathworks.com/help/vision/ref/objectdetectortrainingdata.html