S3OD: Towards Generalizable Salient Object Detection with Synthetic Data
This addresses a data bottleneck for researchers and practitioners in computer vision by enabling more generalizable models without costly annotations, though it is incremental as it builds on existing synthetic data and architecture ideas.
The paper tackles the problem of limited generalization in salient object detection due to expensive annotations by introducing a method that uses large-scale synthetic data generation and an ambiguity-aware architecture, achieving 20-50% error reduction in cross-dataset generalization and state-of-the-art performance on benchmarks.
Salient object detection exemplifies data-bounded tasks where expensive pixel-precise annotations force separate model training for related subtasks like DIS and HR-SOD. We present a method that dramatically improves generalization through large-scale synthetic data generation and ambiguity-aware architecture. We introduce S3OD, a dataset of over 139,000 high-resolution images created through our multi-modal diffusion pipeline that extracts labels from diffusion and DINO-v3 features. The iterative generation framework prioritizes challenging categories based on model performance. We propose a streamlined multi-mask decoder that naturally handles the inherent ambiguity in salient object detection by predicting multiple valid interpretations. Models trained solely on synthetic data achieve 20-50% error reduction in cross-dataset generalization, while fine-tuned versions reach state-of-the-art performance across DIS and HR-SOD benchmarks.