
Zero-shot task-oriented grasping using geometric decomposition and large language models.


coremsg

shapegrasp-zero-shot-task-oriented-grasping-with-large-language-models-through-geometric-decomposition


ShapeGrasp: Zero-Shot Task-Oriented Grasping with Large Language Models through Geometric Decomposition


title_rewrite


RTAGrasp efficiently extracts complete task-oriented grasping constraints from human demonstration videos and transfers them to robots for execution, outperforming existing training-based and training-free methods.


efficient-task-oriented-grasping-through-retrieval-transfer-and-alignment-of-human-demonstration-experiences


Efficient Task-Oriented Grasping through Retrieval, Transfer, and Alignment of Human Demonstration Experiences
