Dataset_BeringSea.zip

File labeled 'Dataset_BeringSea.zip' is the minimal underlying in situ plankton dataset of Bering Sea used for the validation of enhanced CNN proposed by us. All the raw images were acquired by ZOOVIS in the southeastern Bering Sea in May 2017.This dataset contains a training set and a testing set, and both of them contain 7 classes (chaetognatha, copepoda, medusae, euphausiids, fish larvae, limacina and other). In addition, there are 2048 samples in the training set and 512 samples in the testing set for each class. And other files are the data used to validate the Enhanced CNN model proposed by us.

标注为"Dataset_BeringSea.zip"的文件，是用于验证本团队提出的增强型卷积神经网络（Convolutional Neural Network，CNN）的白令海原位浮游生物最小基准数据集。所有原始图像均于2017年5月在白令海东南部由ZOOVIS采集获取。该数据集包含训练集与测试集，二者均涵盖7个类别：毛颚类（chaetognatha）、桡足类（copepoda）、水母类（medusae）、磷虾类（euphausiids）、鱼类幼体（fish larvae）、翼足类（limacina）及其他类群。此外，每个类别的训练集包含2048个样本，测试集包含512个样本。其余文件均为用于验证本团队提出的增强型卷积神经网络模型的验证数据。