PR-MaGIC：通过掩码解码器梯度流进行提示词精炼的上下文分割
PR-MaGIC: Prompt Refinement Via Mask Decoder Gradient Flow For In-Context Segmentation
提出一种无需训练的测试时提示词精炼框架，利用 SAM 掩码解码器的梯度流优化提示词，可无缝集成到现有上下文分割方法中，提升分割质量。
原文依据 PR-MaGIC (Prompt Refinement via Mask Decoder Gradient Flow for In-Context Segmentation), a training-free test-time framework that refines prompts via gradient flow derived from SAM's mask decoder.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论视觉基础模型（VFM）如分割一切模型（SAM）已显著推进图像分割的广泛应用。然而，SAM 及其变体需要大量人工 effort 生成提示词，并针对特定应用进行额外训练。近期方法通过将 SAM 集成到上下文（单样本/少样本）分割中来解决这些限制，利用查询图像与支持图像之间的语义对齐实现自动提示。尽管有这些努力，由于支持图像与查询图像之间的视觉不一致，它们仍会生成次优提示词，从而降低分割质量。为解决这一限制，我们提出 PR-MaGIC（通过掩码解码器梯度流进行提示词精炼的上下文分割），这是一个无需训练的测试时框架，通过源自 SAM 掩码解码器的梯度流来精炼提示词。PR-MaGIC 可无缝集成到上下文分割框架中，具有理论依据，并通过简单的 top-1 选择策略实现实际稳定，确保跨样本的稳健性能。大量评估表明，PR-MaGIC 在各种基准上持续提升分割质量，有效缓解提示词不足的问题，且无需额外训练或架构修改。
英文原文摘要 1313 字符
Visual Foundation Models (VFMs) such as the Segment Anything Model (SAM) have significantly advanced broad use of image segmentation. However, SAM and its variants necessitate substantial manual effort for prompt generation and additional training for specific applications. Recent approaches address these limitations by integrating SAM into in-context (one/few shot) segmentation, enabling auto-prompting through semantic alignment between query and support images. Despite these efforts, they still generate sub-optimal prompts that degrade segmentation quality due to visual inconsistencies between support and query images. To tackle this limitation, we introduce PR-MaGIC (Prompt Refinement via Mask Decoder Gradient Flow for In-Context Segmentation), a training-free test-time framework that refines prompts via gradient flow derived from SAM's mask decoder. PR-MaGIC seamlessly integrates into in-context segmentation frameworks, being theoretically grounded yet practically stabilized through a simple top-1 selection strategy that ensures robust performance across samples. Extensive evaluations demonstrate that PR-MaGIC consistently improves segmentation quality across various benchmarks, effectively mitigating inadequate prompts without requiring additional training or architectural modifications.