OPA 物体放置评估数据集

OPA 全称 Object Placement Assessment，是一个基于 COCO 数据集的物体放置数据集，提供了大量合成图及其前景物体放置合理性二值标签，覆盖 47 个前景种类，共计 73,470 张合成图。正合成图中前景物体的放置合理，而负合成图中前景物体的放置不合理。该数据集也提供了构成合成图的前景物体和背景图片。

OPA, whose full name is Object Placement Assessment, is an object placement dataset based on the COCO dataset. It provides a large number of synthetic images along with their binary labels for evaluating the rationality of foreground object placement, covering 47 foreground categories and totaling 73,470 synthetic images. Positive synthetic images have reasonably placed foreground objects, while negative ones have unreasonably placed foreground objects. The dataset also provides the foreground objects and background images used to construct these synthetic images.

搜集汇总

背景与挑战

背景概述

OPA（Object Placement Assessment）是一个基于COCO数据集的物体放置评估数据集，包含73,470张合成图像，覆盖47种前景物体，每张图像都标注了前景物体放置是否合理的二值标签。该数据集提供了前景物体和背景图片，用于评估物体放置的合理性。