FlickrLogos-32 logo 检测数据集

FlickrLogos-32 是一个目标检测 benchmark，用于在真实场景中进行 logo 检测。图像均源自自然场景，每个图像包含多个 Object 及 1 个 logo 。数据集同时提供了 logo 检测及分类的相关标注。

FlickrLogos-32 is an object detection benchmark dedicated to logo detection in real-world scenarios. All images are sourced from natural scenes, with each image containing multiple objects and exactly one logo. The dataset also provides corresponding annotations for both logo detection and classification.

搜集汇总

背景与挑战

背景概述

FlickrLogos-32是一个目标检测基准数据集，专注于在自然场景中进行logo检测和分类。该数据集包含来自真实场景的图像，每张图像包含多个对象和一个logo，并提供了详细的检测和分类标注。