WHOI-Plankton
收藏资源简介:
WHOI-Plankton是由伍兹霍尔海洋研究所创建的大规模细粒度视觉识别数据集,用于浮游生物分类。该数据集包含超过340万张专家标注的图像,涵盖70个类别。数据集通过Imaging FlowCytobot(IFCB)系统收集,该系统自2006年起持续采集浮游生物图像。数据集的创建旨在通过自动化分类方法解决手动分类大型图像集的不切实际性,并提高分类准确性。该数据集广泛应用于海洋生态学研究,特别是在浮游生物的分类和生态系统动态分析中。
WHOI-Plankton is a large-scale fine-grained visual recognition dataset created by the Woods Hole Oceanographic Institution for plankton classification. This dataset contains over 3.4 million expert-annotated images spanning 70 categories. It was collected via the Imaging FlowCytobot (IFCB) system, which has been continuously acquiring plankton images since 2006. The dataset was developed to address the impracticality of manually classifying large image datasets through automated classification methods, while also improving classification accuracy. It is widely used in marine ecological research, particularly in plankton classification and ecosystem dynamics analysis.




