2 papers
cs.CV2021
RandCrowns: A Quantitative Metric for Imprecisely Labeled Tree Crown Delineation
Dylan Stewart, Alina Zare, Sergio Marconi +5
Supervised methods for object delineation in remote sensing require labeled ground-truth data. Gathering sufficient high quality ground-truth data is difficult, especially when tar…
cs.CV2021
The Weakly-Labeled Rand Index
Dylan Stewart, Anna Hampton, Alina Zare +2
Synthetic Aperture Sonar (SAS) surveys produce imagery with large regions of transition between seabed types. Due to these regions, it is difficult to label and segment the imagery…