发现、分割与选择：面向零样本伪装物体分割的渐进式机制
Discover, Segment, and Select: A Progressive Mechanism for Zero-shot Camouflaged Object Segmentation
把「发现-分割」两阶段换成「发现-分割-选择」三阶段：用视觉特征而非 MLLM 生成候选，再用 MLLM 做语义打分选掩码，这一「候选生成与语义筛选解耦」的思路可直接迁移到开放词表分割、指代分割等需要从多候选中挑最优掩码的任务。
原文依据 a Semantic-driven Mask Selection (SMS) module that employs MLLMs to evaluate and select the optimal segmentation mask from multiple candidates
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论当前的零样本伪装物体分割方法通常采用两阶段流程（先发现后分割）：先使用多模态大语言模型（MLLM）获取视觉提示，再由 SAM 进行分割。然而，仅依赖 MLLM 来发现伪装物体，往往会导致定位不准确、误检和漏检。为解决这些问题，我们提出了发现-分割-选择（DSS）机制，这是一个渐进式细化分割过程的三阶段框架。所提方法包含一个特征驱动的物体发现（FOD）模块，利用视觉特征生成多样化的物体候选；一个分割模块，通过 SAM 分割对这些候选进行细化；以及一个语义驱动的掩码选择（SMS）模块，采用 MLLM 从多个候选中评估并选出最优分割掩码。在四个基准上的大量实验表明，我们的方法在更低的 GPU 显存消耗下取得了最优性能。代码将发布于 https://github.com/ynulonger/DSS.git。
英文原文摘要 1067 字符
Current zero-shot camouflaged object segmentation methods typically employ a two-stage pipeline (discover-then-segment): using MLLMs to obtain visual prompts, followed by SAM segmentation. However, relying solely on MLLMs for camouflaged object discovery often leads to inaccurate localization, false positives, and missed detections. To address these issues, we propose the Discover-Segment-Select (DSS) mechanism, a three-stage framework that progressively refines the segmentation process. The proposed method contains a Feature-driven Object Discovery (FOD) module that leverages visual features to generate diverse object proposals, a segmentation module that refines these proposals through SAM segmentation, and a Semantic-driven Mask Selection (SMS) module that employs MLLMs to evaluate and select the optimal segmentation mask from multiple candidates. Extensive experiments on four benchmarks demonstrate that our method achieves state-of-the-art performance with lower GPU memory consumption. Code will be released at https://github.com/ynulonger/DSS.git.