It detects Koeran traffic roadsigns on image and create data satisfying YOLO's training data format.
It is kind of labeler for refining or creating ground truth on image. It targets to satisfy YOLO's training data foramt.
Diversify the training data for cascade classifying by change the brightness.
Save video frames at fixed fps.
Generate sequence data and empty {image name}.txt, train.txt files
Count the number of labels in dataset
Change the bounding boxes size by same ratio