MuSHRoom
收藏资源简介:
MuSHRoom数据集由坦佩雷大学等机构创建,专注于室内房间规模的场景,包含10个真实世界房间的数据。每个房间使用Kinect和iPhone等消费级设备捕捉RGB-D序列,并使用Faro扫描仪获取精确的地面实况网格模型。数据集旨在为房间规模的3D重建和新视角合成提供基准,解决现有数据集在几何重建和照片级真实感建模方面的不足。MuSHRoom数据集通过模拟VR/AR应用场景,提出了包括遮挡、运动模糊、反射、透明度和光照变化等实际挑战,推动了在消费级设备上进行鲁棒和沉浸式建模与渲染技术的发展。
The MuSHRoom dataset was developed by Tampere University and other research institutions, focusing on indoor room-scale scenarios, and comprises data collected from 10 real-world rooms. For each room, RGB-D sequences were captured using consumer-grade devices such as Kinect and iPhone, while precise ground-truth mesh models were acquired via a Faro scanner. The dataset is designed as a benchmark for room-scale 3D reconstruction and novel view synthesis, addressing the shortcomings of existing datasets in geometric reconstruction and photorealistic modeling. By simulating VR/AR application scenarios, the MuSHRoom dataset introduces practical challenges including occlusion, motion blur, reflection, transparency and illumination variations, promoting the development of robust and immersive modeling and rendering technologies on consumer-grade devices.




