MatPredict
收藏资源简介:
MatPredict数据集是一个用于学习室内物体材料属性的综合数据集。该数据集结合了Replica数据集中的高质量合成物体和MatSynth数据集中的材料属性类别,以创建具有多样材料属性的物体。数据集中包含了18个常见物体和14种不同的材料。此外,数据集还展示了如何在不同的光照和相机位置下为这些物体提供多样性。该数据集旨在为计算机视觉社区在室内机器人应用方面的工作提供支持,并促进室内机器人技术的进步。
The MatPredict Dataset is a comprehensive dataset for learning material properties of indoor objects. It combines high-quality synthetic objects from the Replica dataset and material property categories from the MatSynth dataset to create objects with diverse material attributes. The dataset contains 18 common object categories and 14 distinct material types. Additionally, it provides diversity for these objects under varying lighting conditions and camera poses. This dataset aims to support the work of the computer vision community in indoor robotic applications and promote the advancement of indoor robotics technology.

- 1MatPredict: a dataset and benchmark for learning material properties of diverse indoor objects密歇根大学交通研究所 · 2025年



