PulseAugur
实时 01:15:22
English(EN) AnythingReality: Robust Online Gaussian Splatting SLAM for Open-Vocabulary VR Scene Exploration

AnythingReality 集成高斯溅射、SLAM 和 VLM 以实现交互式 VR

研究人员开发了 AnythingReality,这是一个将 3D 高斯溅射与同步定位与地图构建 (SLAM) 集成到虚拟现实 (VR) 环境中的新颖系统。该系统通过利用 ORB-SLAM3 进行姿态估计,从嘈杂的真实世界数据中实现场景的实时重建,从而增强了 VR 探索。此外,它还集成了视觉语言模型 (VLM),支持语音驱动的交互,允许用户发出语音命令、生成场景描述和标记兴趣点,VLM 对象识别率达到 88%。 AI

影响 通过集成的 AI 功能增强 VR 场景探索和交互,可能改善用户在虚拟环境中的体验。

排序理由 该集群描述了一篇详细介绍新颖技术架构及其性能改进的研究论文。

在 Hugging Face Daily Papers 阅读 →

AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →

AnythingReality 集成高斯溅射、SLAM 和 VLM 以实现交互式 VR

报道来源 [3]

  1. Hugging Face Daily Papers TIER_1 English(EN) ·

    AnythingReality:面向开放词汇VR场景探索的鲁棒在线高斯溅射SLAM

    We present a novel integrated architecture for robust online 3D Gaussian splatting, real-time VR exploration, and speech-driven Vision-Language-Model interaction. Unlike methods assuming clean depth or external poses, our system combines ORB-SLAM3-based pose estimation with onlin…

  2. arXiv cs.CV TIER_1 English(EN) · Timofei Kozlov, Dmitrii Maliukov, Andrey Marchenko, Miguel Altamirano Cabrera, Dzmitry Tsetserukou ·

    AnythingReality: 鲁棒的在线高斯溅射SLAM,用于开放词汇VR场景探索

    arXiv:2607.09260v1 Announce Type: new Abstract: We present a novel integrated architecture for robust online 3D Gaussian splatting, real-time VR exploration, and speech-driven Vision-Language-Model interaction. Unlike methods assuming clean depth or external poses, our system com…

  3. arXiv cs.CV TIER_1 English(EN) · Dzmitry Tsetserukou ·

    AnythingReality:面向开放词汇VR场景探索的鲁棒在线高斯溅射SLAM

    We present a novel integrated architecture for robust online 3D Gaussian splatting, real-time VR exploration, and speech-driven Vision-Language-Model interaction. Unlike methods assuming clean depth or external poses, our system combines ORB-SLAM3-based pose estimation with onlin…