Search papers, labs, and topics across Lattice.
This paper introduces a reliability-aware geometry distillation framework for RGB-D salient object detection (SOD) that operates without relying on potentially unreliable depth data during training or inference. By utilizing a frozen Depth Anything V2 model as a teacher to transfer essential geometric information and employing a pixel-wise reliability estimator, the method effectively enhances the performance of RGB-only detectors. The proposed approach achieves state-of-the-art results in 26 out of 36 comparisons against existing RGB-D methods, demonstrating a significant reduction in mean absolute error and improved performance across different datasets.
RGB-only salient object detection can outperform traditional RGB-D methods by leveraging reliable geometry distillation without using depth data during inference.
Depth can resolve appearance ambiguity in RGB-D salient object detection (SOD), yet sensor depth is not uniformly reliable. Missing regions, blurred boundaries, and structural artifacts can propagate through multimodal fusion and make an RGB-D detector less accurate than its RGB-only counterpart. Existing quality-aware approaches regulate observed depth but remain dependent on the same potentially defective modality. We propose \method, a reliability-aware geometry distillation framework developed for RGB-D SOD benchmarks without using dataset-provided depth during training or inference. A frozen Depth Anything V2 model serves only as a training-time teacher, transferring dense relative geometry, hierarchical spatial attention, and boundary structure to a compact edge-aware geometry branch. Pooled bidirectional interaction aligns geometry with appearance, and a pixel-wise reliability estimator selectively injects geometry that is compatible with the current RGB representation. The teacher is removed after training, leaving an RGB-only inference network. Trained on 2,985 RGB-mask pairs, \method{} achieves the best or tied-best result in 26 of 36 metric-dataset comparisons against ten recent RGB-D SOD methods, including a 13.4\% relative MAE reduction on ReDWeb-S. When retrained on DUTS-TR, it also improves the strongest prior $F$-measure by 4.2\% on PASCAL-S, showing that the distilled geometry transfers beyond a particular sensor or dataset domain. Code will be released upon publication.
