logo-matching 商标识别数据集

logo-matching 涵盖了超过 200 个不同品牌和机构的标志图像和标签。每个图像均经过精确标注，确保了数据的高质量；标签是一个序列，包含 208 个类别的名称，涵盖了多个领域的实体，如公司、组织、媒体等。数据集分为训练集、验证集和测试集，分别包含 1077 、 248 和 258 个样本。

The logo-matching dataset contains logo images and corresponding labels for over 200 distinct brands and institutions. Each image has been meticulously annotated to ensure high data quality; the labels are a sequence comprising 208 category names, covering entities across various domains such as companies, organizations, media outlets, and more. The dataset is split into training, validation, and test sets, with 1077, 248, and 258 samples respectively.

搜集汇总

背景与挑战

背景概述

logo-matching商标识别数据集包含超过200个品牌和机构的标志图像，每个图像都经过精确标注，涵盖208个类别。该数据集分为训练集、验证集和测试集，分别有1077、248和258个样本，适用于图像分类和品牌识别等任务。