DAQUAR 真实世界图像问答数据集

DAQUAR，全称 DAtaset for QUestion Answering on Real-world images，是一个关于图像的人类问答的数据集。该数据集的图像来自于 NYU-Depth v2 数据集，都是关于室内场景的 RGBD 图像，其中 795 张用于训练，654 张用于测试。 DAQUAR 中的 question/answer pairs 主要有两种类型：自动生成的和人工标注的。

DAQUAR, whose full name is Dataset for Question Answering on Real-world Images, is a dataset dedicated to human question answering related to images. The images in this dataset are sourced from the NYU-Depth v2 dataset, all of which are RGBD images of indoor scenes, with 795 samples allocated for training and 654 for testing. There are two primary types of question/answer pairs in DAQUAR: automatically generated pairs and manually annotated pairs.

搜集汇总

背景与挑战

背景概述

DAQUAR是一个关于真实世界图像问答的数据集，包含来自NYU-Depth v2数据集的室内场景RGBD图像，共有795张训练图像和654张测试图像。数据集提供自动生成和人工标注的问答对，主要用于目标检测和图像理解任务。