TF-SSD：通过协同掩码筛选构建免训练协同显著性目标检测的强流水线
TF-SSD: A Strong Pipeline via Synergic Mask Filter for Training-free Co-salient Object Detection
展示如何免训练地协同 SAM 与 DINO：用 SAM 生成候选掩码池，再用 DINO 注意力图做图像内显著性筛选、用跨图像原型相似度做图像间选择，可直接迁移到其他需要跨图像一致性的分割任务。
原文依据 we propose a novel training-free method, TF-SSD, through the synergy between SAM and DINO
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论协同显著性目标检测（CoSOD）旨在分割在一组相关图像中持续出现的显著目标。尽管近期基于训练的方法取得了显著进展，但它们仍受限于闭集数据集，泛化能力有限。然而，很少有研究探索视觉基础模型（VFM）解决 CoSOD 的潜力，这些模型展现出强大的泛化能力和稳健的显著性理解。本文研究并利用 VFM 进行 CoSOD，进一步提出一种新颖的免训练方法 TF-SSD，通过 SAM 与 DINO 的协同实现。具体而言，我们首先利用 SAM 生成全面的原始候选，作为候选掩码池。然后，引入质量掩码生成器过滤冗余掩码，从而获得精炼掩码集。由于该生成器基于 SAM 构建，其本身缺乏对显著性的语义理解。为此，我们采用图像内显著性过滤器，利用 DINO 的注意力图识别单张图像中视觉显著的掩码。此外，为将显著性理解扩展到组图像之间，我们提出图像间原型选择器，计算跨图像原型之间的相似度分数，以选择得分最高的掩码。这些选定的掩码作为 CoSOD 的最终预测。大量实验表明，我们的 TF-SSD 优于现有方法（例如，相比近期免训练方法提升 13.7%）。代码见 https://github.com/hzz-yy/TF-SSD。
英文原文摘要 1561 字符
Co-salient Object Detection (CoSOD) aims to segment salient objects that consistently appear across a group of related images. Despite the notable progress achieved by recent training-based approaches, they still remain constrained by the closed-set datasets and exhibit limited generalization. However, few studies explore the potential of Vision Foundation Models (VFMs) to address CoSOD, which demonstrate a strong generalized ability and robust saliency understanding. In this paper, we investigate and leverage VFMs for CoSOD, and further propose a novel training-free method, TF-SSD, through the synergy between SAM and DINO. Specifically, we first utilize SAM to generate comprehensive raw proposals, which serve as a candidate mask pool. Then, we introduce a quality mask generator to filter out redundant masks, thereby acquiring a refined mask set. Since this generator is built upon SAM, it inherently lacks semantic understanding of saliency. To this end, we adopt an intra-image saliency filter that employs DINO's attention maps to identify visually salient masks within individual images. Moreover, to extend saliency understanding across group images, we propose an inter-image prototype selector, which computes similarity scores among cross-image prototypes to select masks with the highest score. These selected masks serve as final predictions for CoSOD. Extensive experiments show that our TF-SSD outperforms existing methods (e.g., 13.7% gains over the recent training-free method). Codes are available at https://github.com/hzz-yy/TF-SSD.