UAVS model trained on all classes collapsed to one class. No Batray localizations included
Here are the input classes (labels) that were collapsed to one class, "object"
'Bird','Boat','Buoy','Egregia','Foam','Jelly','Kelp','Mola','Mooring','Otter','Person','Pinniped','Excretia','RIB','Reflectance','Secci_Disc','Shark','Surfboard','Wave','Whale','Wood'
input to model:
Input images: 1226; transformed images: 26928.
Input labels: {'object': 27757}; transformed labels: {'object': 105871}
Validation run (10% of labels)
image/png
image/png
image/png
image/png
image/png
image/png
image/png
image/png
