Step 2 in the setup:
Download and extract augmented segmentation data (Thanks to DrSleep),
I downloaded this data, and looking at it, it seems to only have the outlines of the objects in white.
What is the purpose of this augmented data? We already have the labels from VOC.. im not sure how the network can learn from this augmented data.
I am following the instructions in the README.
Step 2 in the setup: Download and extract augmented segmentation data (Thanks to DrSleep),
I downloaded this data, and looking at it, it seems to only have the outlines of the objects in white. What is the purpose of this augmented data? We already have the labels from VOC.. im not sure how the network can learn from this augmented data.