Benthic substrate classification ML training image snips

Collection of images labelled into substrate classes for training machine learning algorithms. \nLineage: Benthic Stereo still imagery was collected during the MNF Investigator voyage IN2018_V06, sampling seamounts off Tasmania. Random points (5per m2) were annotated within measured quadrats for substrate classes including matrix-forming stony coral (Metadata description in Marlin). Snips centred on these random points were extracted from the original images, and associcated with their relevant substrate class label.