EEG-EditBench：通过可控图像编辑探测EEG-图像检索模型中的视觉信息
EEG-EditBench: Probing Visual Information in EEG-Image Retrieval Models with Controlled Image Edits
AI总结：
该研究针对EEG-图像检索模型无法揭示支撑匹配的视觉信息这一问题，构建EEG-EditBench诊断基准，评估8个代表性模型，发现细粒度属性变化是最大挑战，为相关研究提供可控基础。
AI中文摘要：
近期的EEG-图像检索模型在从语义多样的候选图像中识别被观看图像方面已取得优异性能，但这类成功并未揭示支撑匹配的视觉信息是什么。模型或许能轻易在工具、植物和车辆中识别出猎豹，但它能否将被观看的猎豹与将猎豹替换为狗的同一场景区分开？受该问题驱动，我们推出EEG-EditBench，这是一个诊断基准，通过对物体身份、属性、背景及物体存在进行可控编辑来探究上述问题。EEG-EditBench基于200 THINGS-EEG2测试图像构建，包含2137个经质量控制的编辑，并评估8个代表性EEG视觉解码模型。我们的结果表明，优异的标准检索性能并不能一致地迁移到基于编辑的评估中，细粒度属性变化构成最大挑战。EEG-EditBench揭示了被整体检索准确率所掩盖的模型行为，为研究EEG-图像模型保留了哪些视觉信息提供了可控基础，代码和完整数据集已公开可用。
英文摘要：
Recent EEG-to-image retrieval models have achieved strong performance in identifying viewed images from semantically diverse candidates. Yet such success does not reveal what visual information supports the match. A model may readily identify a cheetah among tools, plants, and vehicles, but can it still distinguish the viewed cheetah from the same scene with the cheetah replaced by a dog? Motivated by this question, we introduce EEG-EditBench, a diagnostic benchmark that examines this question through controlled edits of object identity, attributes, background, and object presence. Built from the 200 THINGS-EEG2 test images, EEG-EditBench contains 2,137 quality-controlled edits and evaluates eight representative EEG visual decoding models. Our results show that strong standard retrieval does not consistently transfer to edit-based evaluation, with fine-grained attribute changes presenting the greatest challenge. EEG-EditBench reveals model behavior hidden by aggregate retrieval accuracy and provides a controlled basis for studying what visual information EEG-image models preserve. The code and complete dataset are publicly available.
