A Pyramid CNN for Dense-Leaves Segmentation
This work addresses leaf segmentation in natural environments, which is incremental as it builds on existing CNN and watershed methods with a new dataset.
The authors tackled the problem of segmenting overlapping leaves in dense foliage by introducing the Dense-Leaves dataset and a pyramid CNN with multi-scale predictions for boundary detection, combined with a watershed algorithm for instance segmentation, achieving promising results.
Automatic detection and segmentation of overlapping leaves in dense foliage can be a difficult task, particularly for leaves with strong textures and high occlusions. We present Dense-Leaves, an image dataset with ground truth segmentation labels that can be used to train and quantify algorithms for leaf segmentation in the wild. We also propose a pyramid convolutional neural network with multi-scale predictions that detects and discriminates leaf boundaries from interior textures. Using these detected boundaries, closed-contour boundaries around individual leaves are estimated with a watershed-based algorithm. The result is an instance segmenter for dense leaves. Promising segmentation results for leaves in dense foliage are obtained.