Figure 1
From: Explainable few-shot learning workflow for detecting invasive and exotic tree species
Overall workflow of the study, showing the three main processing steps: (1) the pre-training of a base Siamese network using images of tree canopies from existing benchmarks and automatically generated from the UAV image of the study area, (2) refining the networks in a few-shot learning setup, and (3) adding a post-hoc explanation to visually inform users whether the network can distinguish the new tree species.
