Add n/labels and n/images subfolders to output
Currently, the merge_and_segment_training_data.py
script outputs into folders 0, 1, 2, ... N directly, with no subfolders.
Yolo expects a file path containing images
for training files. It also expects that it may substitute images
in its filepath for labels
to find the location of labels (in yolo .txt format) for each respective image. For small numbers of N, this is easy to do by hand, but it should probably be made automatic before we publish this script
The problem that this script solves may be very transferable to many other ML practitioners working on similar problems. If we clean up this script a little bit and publish it, it may bring recognition to our team
See also (candidates for open source publication):
https://gitlab.com/KSU_EVT/autonomous-software/get-orange-cones
https://gitlab.com/KSU_EVT/autonomous-software/obtain_training_images