Replica

该数据集名为Replica，包含了18个不同的室内场景，如办公室、酒店和公寓等，为室内空间提供了高质量和逼真的重建。这些场景被划分为互斥的训练、验证和测试集。为了提升物体检测的效果，该方法生成了2D和3D伪标签。数据集规模包括18个独特场景，其中训练集有17,000张图像，验证集有1,000张，测试集有2,300张。任务内容包括2D和3D物体检测。

This dataset, named Replica, includes 18 distinct indoor scenes such as offices, hotels, apartments, etc., providing high-quality and photorealistic reconstructions of indoor spaces. These scenes are divided into mutually exclusive training, validation and test sets. To enhance object detection performance, 2D and 3D pseudo-labels are generated for this dataset. The dataset has 17,000 training images, 1,000 validation images, and 2,300 test images across its 18 unique scenes. The supported tasks include 2D and 3D object detection.

搜集汇总

背景与挑战

背景概述

Replica数据集是一个高质量的室内空间重建数据集，包含18个场景的密集几何、高分辨率纹理和丰富的语义信息，支持AI Habitat框架，适用于机器学习和计算机视觉研究。