SAGE3D: Soft-guided attention and graph excitation for 3D point cloud corner detection
This work addresses the problem of accurate corner detection in 3D point clouds for remote sensing applications, offering a novel approach that enhances both precision and recall.
SAGE3D introduces a hybrid Transformer-based model for corner detection in airborne LiDAR point clouds, achieving improved precision and recall through soft-guided attention and excitatory graph neural networks. The method demonstrates state-of-the-art performance on benchmark datasets.
We present SAGE3D, a hybrid Transformer-based model for corner detection in airborne LiDAR point clouds. We propose a multi-stage solution built on a hierarchical encoder-decoder architecture that progressively downsamples point clouds through Set Abstraction layers and recovers per-point predictions via Feature Propagation. We introduce two innovations: Soft-Guided Attention, which injects ground-truth corner labels as a log-prior into attention logits during training to improve precision; then an Excitatory Graph Neural Network positioned at strategic resolutions in the hierarchy, employing positive-only message passing where high-confidence corners reinforce predictions through learned boosting, optimizing for recall. The hierarchical design enables multi-scale feature extraction while our guided attention and excitatory modules ensure corner signals are amplified rather than diluted across scales.