SceneVerse:百万级别的3D视觉语言数据集,3D场景理解新SOTA
“SCENEVERSE: Scaling 3D Vision-Language Learning for Grounded Scene Understanding”
项目主页:https://scene-verse.github.io
论文地址:https://arxiv.org/pdf/2...
GenZI:零样本3D人体-场景交互生成方法
“GenZI: Zero-Shot 3D Human-Scene Interaction Generation”
给定任意 3D 场景,GenZI 利用视觉语言模型(VLM)的强大能力,可以根据简...