An ice-water classification dataset for learning from label proportions 2015 [Canada]

An ice-water classification dataset for machine learning researchers to test algorithms related to learning from label proportions. The dataset includes patches extracted from 10 RADARSAT-2 dual-polarized scenes, Egg Code polygon (bag) IDs, labels, and the estimated proportion of each bag provided b...

Full description

Bibliographic Details
Main Authors: Li, Fan, Taylor, Graham
Format: Other/Unknown Material
Language:unknown
Published: Borealis 2015
Subjects:
Engineering
classification
label proportion
sea ice
weakly-supervised learning
Canada
Sea ice
Online Access:https://hdl.handle.net/10864/WJ7VY
Description
Summary:An ice-water classification dataset for machine learning researchers to test algorithms related to learning from label proportions. The dataset includes patches extracted from 10 RADARSAT-2 dual-polarized scenes, Egg Code polygon (bag) IDs, labels, and the estimated proportion of each bag provided by ice experts.

Similar Items