CUP- Convolutional Neural Network Training dataset

The dataset has been provided for training a convolutional neural network to measure some selected visual design principles. These visual principles are related to the preference matrix which is adopted from Kaplan and Kaplan (1989). The dataset is contained of images that illustrate the considered variables as obviously as possible and the CNN trained by them can be used for analysis in fields of art and architecture. CUP is the short form of contrast, unity, and proportion.
3 types of complementary color contrasts, 2 types of warm/cold contrasts, light/dark contrast, similarity in color and variety in form , proportion and similarity in form and variety in color.
The dataset is a collection of images found in the search through google and pinterest, and some of them may be subject to copyright. For such images, the copyright of all the images belongs to the image owners.

本数据集用于训练卷积神经网络（Convolutional Neural Network, CNN），以量化若干选定的视觉设计原则。这些视觉原则与源自Kaplan与Kaplan（1989）提出的偏好矩阵相关联。
本数据集由尽可能清晰地呈现所考量变量的图像组成，经该数据集训练得到的卷积神经网络可应用于艺术与建筑领域的分析工作。CUP为对比度、统一性与比例的英文缩写。
具体涵盖三类互补色对比度、两类冷暖色调对比度、明暗对比度，以及色彩相似性与形式多样性、比例关系、形式相似性与色彩多样性等视觉要素。
本数据集的图像均通过谷歌（Google）与Pinterest平台检索获取，其中部分图像可能受版权保护，所有图像的版权均归原图像所有者所有。