MLLM零样本语言推理用于跨视角地理定位研究

1. 12:00
   arXiv · Computer Vision

   What Words Keep of a Place：跨视角地理定位的零样本语言推理研究

   研究提出零样本跨视角地理定位方法，用多模态大语言模型（MLLM）把地面全景与卫星瓦片描述成结构化文本并比对，全程无需训练。在 9,826 对 VIGOR 数据上，全库按描述相似度排序 Recall@1 仅 0.39%；缩到 10 个邻近瓦片后，MLLM 评判器将随机排序效果翻倍并追平强词法基线，还能指出字段一致或冲突。

Current heat 9·Comparable peak 10（Oct 7）·Comparable change over 24 hours –

The trend compares only the same participants observed continuously; its range may be smaller than the current heat count. Move or click on the chart to inspect hourly heat; use the left and right arrow keys to switch.