See like a Robot: Robot-Centric Pointmaps for Vision-Language-Action Models

  • 类型:arxiv
  • 标识:2607.11498
  • 链接:https://arxiv.org/abs/2607.11498
  • 主分类:multimodal
  • 形态:method
  • 被引:0
  • 被引来源:Semantic Scholar + OpenAlex
  • S2被引:0
  • OpenAlex被引:0
  • 影响力被引:0
  • TLDR:SeeR-VLA is proposed, which transforms pointmaps into a robot-centric frame with an end-effector origin and robot-base-aligned axes, and an encoder initialized from pretrained RGB weights extracts pointmap features, which are added to corresponding RGB tokens without increasing the token count.
  • OpenAlex ID:W7168294455
  • OpenAlex DOI:10.48550/arxiv.2607.11498
  • DOI:10.48550/arxiv.2607.11498
  • DOI来源:OpenAlex
  • 开放获取:green
  • 开放获取链接:https://doi.org/10.48550/arxiv.2607.11498
  • OpenAlex更新:2026-09-03
  • 待LLM分类:否
  • 标题中文:像机器人一样看:面向视觉-语言-动作模型的机器人中心点图
  • TLDR中文:SeeR-VLA 将点图变换到以末端执行器为原点、机器人基座轴对齐的机器人中心坐标系,并由基于预训练 RGB 权重初始化的编码器提取点图特征,将其在不增加 token 数的情况下叠加到对应的 RGB token 上。
  • 来源文件:
  • /inbox/tom/_candidates/2026-07-21-rag-retrieval-reranking-candidates.json
  • /inbox/tom/_candidates/2026-07-21-agent-rag-longcontext-candidates.json
  • [S2 enrich]
  • [OpenAlex backfill]