graviti
产品服务
解决方案
知识库
公开数据集
关于我们
Architectural Style Facade
2D Box
Urban
|...
许可协议: Unknown

Overview

The Architectural Style Facade dataset is a 25-class dataset used for architectural style classification and regression.(The download link has been updated 16/09/2014).downloadlink (google) or downloadlink (baidu)The ten styles used in our first experiments are: (randomly 30 images used for training and 10-fold cross validation)* American craftsman style* Baroque architecture* Chicago school architecture* Colonial architecture* Georgian architecture* Gothic architecture* Greek Revival architecture* Queen Anne architecture* Romanesque architecture* Russian Revival architectureFor a better evaluation of inter-class relationships between architectural styles, we are further collecting a larger number of styles (for now 66 styles), and tries to manage more metadata such as the name and place of the buildings. Please keep in track if you are interested in the task.This dataset contains approximately 5000 images from 25 architectural styles. The raw images are downloaded from Wikimedia Commons with the first 5000 images from each style. We manually filtered the images to exclude images of interior decoration, building parts, or severe scale and orientation change, and resulted in a total of 5000 images.For reference, the time period and style relationships between styles checked with Wikipedia are provided in the txt files.Some of the images are collected from the 4-class dataset, reported in the paper:Visual pattern discovery for architecture image classification and product image searchPlease cite the paper if you use the dataset:Architectural Style Classification using Multinomial Latent Logistic Regression (ECCV2014)Feel free to contact Zhe Xu if you have some questions or advice to the dataset.http://commons.wikimedia.org/wiki/Category:Architecture_by_stylehttps://drive.google.com/file/d/0Bwo0SFiZwl3JVGRlWGZUaW5va00/edit?usp=sharinghttp://pan.baidu.com/s/1qWqj9BA

数据概要
数据格式
image,
数据量
--
文件大小
--
| 数据量 -- | 大小 --
Architectural Style Facade
2D Box
Urban
许可协议: Unknown

Overview

The Architectural Style Facade dataset is a 25-class dataset used for architectural style classification and regression.(The download link has been updated 16/09/2014).downloadlink (google) or downloadlink (baidu)The ten styles used in our first experiments are: (randomly 30 images used for training and 10-fold cross validation)* American craftsman style* Baroque architecture* Chicago school architecture* Colonial architecture* Georgian architecture* Gothic architecture* Greek Revival architecture* Queen Anne architecture* Romanesque architecture* Russian Revival architectureFor a better evaluation of inter-class relationships between architectural styles, we are further collecting a larger number of styles (for now 66 styles), and tries to manage more metadata such as the name and place of the buildings. Please keep in track if you are interested in the task.This dataset contains approximately 5000 images from 25 architectural styles. The raw images are downloaded from Wikimedia Commons with the first 5000 images from each style. We manually filtered the images to exclude images of interior decoration, building parts, or severe scale and orientation change, and resulted in a total of 5000 images.For reference, the time period and style relationships between styles checked with Wikipedia are provided in the txt files.Some of the images are collected from the 4-class dataset, reported in the paper:Visual pattern discovery for architecture image classification and product image searchPlease cite the paper if you use the dataset:Architectural Style Classification using Multinomial Latent Logistic Regression (ECCV2014)Feel free to contact Zhe Xu if you have some questions or advice to the dataset.http://commons.wikimedia.org/wiki/Category:Architecture_by_stylehttps://drive.google.com/file/d/0Bwo0SFiZwl3JVGRlWGZUaW5va00/edit?usp=sharinghttp://pan.baidu.com/s/1qWqj9BA

0
立即开始构建AI
graviti
wechat-QR
长按保存识别二维码,关注Graviti公众号

Copyright@Graviti
沪ICP备19019574号
沪公网安备 31011002004865号