RoboSeg
收藏资源简介:
RoboSeg数据集由清华大学交叉信息研究院、清华大学计算机科学与技术系、上海琦智研究院和上海人工智能实验室共同创建,包含3800张针对机器人场景进行高质量分割注释的图像。该数据集覆盖了多种类型的机器人、摄像头视角和背景环境,为训练首个通用的、高质量的机器人分割模型提供了基础。通过RoboSeg数据集训练出的模型,能够实现无需校准和即插即用的机器人分割与增强技术。
The RoboSeg dataset was jointly created by the Institute for Interdisciplinary Information Sciences at Tsinghua University, Department of Computer Science and Technology of Tsinghua University, Shanghai Qizhi Institute, and Shanghai AI Laboratory. It contains 3800 high-quality images with fine segmentation annotations tailored for robotic scenarios. This dataset covers diverse robot types, camera perspectives and background environments, laying a solid foundation for training the first universal, high-quality robotic segmentation model. Models trained on the RoboSeg dataset can achieve calibration-free and plug-and-play robotic segmentation and enhancement technologies.




