DIODE 单目深度估计数据集

DIODE 全称 Dense Indoor and Outdoor DEpth，是首个用于单目深度估计的数据集。该数据集包括了由相同硬件设置获得的不同室内与室外场景。该数据集的训练集包括了 8574 个室内和 16884 个室外样本。验证集包含了 325 个室内和 446 个室外样本。

DIODE, which stands for Dense Indoor and Outdoor DEpth, is the first dataset dedicated to monocular depth estimation. This dataset includes diverse indoor and outdoor scenes acquired with identical hardware configurations. The training split of this dataset consists of 8,574 indoor samples and 16,884 outdoor samples, while the validation split contains 325 indoor samples and 446 outdoor samples.

搜集汇总

背景与挑战

背景概述

DIODE数据集是首个专为单目深度估计设计的数据集，涵盖了室内和室外场景，由相同硬件采集。其训练集包含8574个室内样本和16884个室外样本，验证集则包括325个室内样本和446个室外样本。