Meta-Album
收藏arXiv2023-02-16 更新2024-08-06 收录
下载链接：
http://arxiv.org/abs/2302.08909v1
下载链接
链接失效反馈官方服务：
资源简介：
Meta-Album是一个图像分类元数据集，旨在促进少样本学习、迁移学习、元学习等任务。它包含40个开放数据集，每个数据集至少有20个类别，每个类别有40个示例，具有经过验证的许可证。这些数据集来自生态学（动物和植物）、制造业（纹理、车辆）、人类行为和光学字符识别等多样领域，具有不同的图像尺度（显微镜、人类尺度、遥感）。所有数据集都经过预处理、标注和统一格式化，并提供3个版本（Micro ⊆ Mini ⊆ Extended）以匹配用户的计算资源。
Meta-Album is an image classification meta-dataset designed to facilitate few-shot learning, transfer learning, meta-learning and other related tasks. It comprises 40 open datasets, each containing at least 20 categories with 40 examples per category and validated licenses. These datasets cover diverse domains including ecology (animals and plants), manufacturing (textures, vehicles), human behavior, optical character recognition (OCR) and more, with varying image scales (microscopic, human-scale, remote sensing). All datasets have been preprocessed, annotated and uniformly formatted, and three versions (Micro ⊆ Mini ⊆ Extended) are provided to accommodate users' computational resources.
创建时间：
2023-02-16
