Deep neural network feature maps

Feature maps summaryHere we release the PCA-downsampled deep neural network (DNN) feature maps used in the data resource paper: "A large and rich EEG dataset for modeling human visual object recognition". We used four DNN architectures (AlexNet, ResNet-50, CORnet-S, MoCo), and extracted their feature map responses to images coming from the THINGS database and from the ILSVRC-2012 challenge.Useful materialAdditional informationFor additional information on the DNNs used, the stimuli images and feature maps extraction procedure please refer to our paper and code.Additional dataset resourcesPlease visit the dataset page for the paper, dataset tutorial, code and more.OSFFor additional data and resources visit our OSF project, where you can find:The stimuli imagesA detailed descriptions of the DNN feature maps data filesCitationsIf you use any of our data, please cite our paper.

特征图汇总
本团队发布了数据资源论文《用于人类视觉物体识别建模的大型丰富脑电图（Electroencephalogram, EEG）数据集》中所使用的经主成分分析（Principal Component Analysis, PCA）降采样的深度神经网络（Deep Neural Network, DNN）特征图。本研究采用四种DNN架构（AlexNet、ResNet-50、CORnet-S、MoCo），提取了这些架构对来自THINGS数据库与ILSVRC-2012挑战赛图像的特征图响应。

实用资料
补充信息
如需了解所用DNN、刺激图像及特征图提取流程的更多细节，请参阅本团队的论文与代码。

附加数据集资源
请访问本论文的数据集页面，获取数据集教程、代码及更多相关内容。

OSF
如需获取更多数据与资源，请访问我们的OSF项目，其中包含：刺激图像、DNN特征图数据文件的详细说明。

引用说明
若您使用本数据集，请引用我们的相关论文。