ArchitecturalStyles

ArchitecturalStyles是2022IAIF数据集，一个用于开发深度学习算法的图像分类数据集。其数据主要来源于Architectural style classification数据集。由24种建筑风格组成，主要格式为ImageNet格式，可适用于初级图像分类算法训练与测试。训练集和验证集在ArchitecturalStyles_train文件夹中，训练集有7486张图片，验证集有1083张，图片均按类别放置，测试集共1169张图片，在ArchitecturalStyles_test文件夹中。

ArchitecturalStyles is the 2022 IAIF dataset, an image classification dataset designed for developing deep learning algorithms. Its data is mainly sourced from the Architectural Style Classification dataset. This dataset comprises 24 architectural styles, and the images mostly follow the ImageNet format, making it suitable for training and testing basic image classification algorithms. The training and validation sets are stored in the ArchitecturalStyles_train folder, where the training set contains 7,486 images and the validation set has 1,083 images, with all images organized by their respective categories. The test set, consisting of 1,169 images in total, is located in the ArchitecturalStyles_test folder.

搜集汇总

背景与挑战

背景概述

ArchitecturalStyles是一个2022年发布的图像分类数据集，专注于24种建筑风格，采用ImageNet格式，适用于初级深度学习算法训练与测试。数据集包含训练集（7486张图片）、验证集（1083张图片）和测试集（1169张图片），总大小1.2GB，由IAIF机构发布，遵循CC BY-NC-SA 4.0许可证。