Abstract: Semi-supervised semantic segmentation with consistency regularization capitalizes on unlabeled images to enhance the accuracy of pixel-level segmentation. Current consistency learning ...

Abstract: Forest tree species classification using remote sensing data often faces the challenge of limited labeled samples, which hampers the performance of deep learning models. While few-shot ...