Herbarium 2021 Half–Earth

植物标本馆 2022：北美植物群是纽约植物园项目的一部分，该项目由美国国家科学基金会资助，旨在构建用于识别世界各地新植物物种的工具。该数据集使用从全球 60 个不同植物机构收集的图像，力求代表北美所有已知的维管植物分类群。
我们的数据集具有长尾分布。每个分类单元的图像数量少则 7 幅，多则 100 幅。尽管有更多图像可用，但我们限制了最大数量，以确保竞赛参与者有足够但可管理的训练数据大小。

Herbarium 2022: North American Flora is part of a project by the New York Botanical Garden, funded by the U.S. National Science Foundation, which aims to develop tools for identifying new plant species across the globe. This dataset utilizes images collected from 60 distinct botanical institutions worldwide, and seeks to represent all known vascular plant taxa in North America.
The dataset exhibits a long-tailed distribution. The number of images per taxon ranges from as few as 7 to as many as 100. While more images were available, we capped the maximum number of images per taxon to ensure that competition participants have sufficient yet manageable training data sizes.

搜集汇总

背景与挑战

背景概述

该数据集是纽约植物园项目的一部分，旨在开发识别全球新植物物种的工具，包含来自全球60个植物机构的图像，覆盖北美所有已知维管植物分类群。其数据呈长尾分布，每个分类单元的图像数量限制在7到100幅之间，以提供适度规模的训练数据。