Previously, I used the bounding box labeling method in Edge Impulse to train an object detection model, and it worked very well.
Currently, I have another project that requires training a model using polygon segmentation labeling. The labeling format is as follows:
1 0.7421875 0.4443059833333333 0.73828125 0.42129984583333335 0.724609375 0.39541794583333334 0.7421875 0.4443059833333333
However, I couldn’t find any place on the Edge Impulse website where I could import my labeled training dataset in this format.
Does Edge Impulse support polygon segmentation labeling?
If not, are there any alternative approaches?