中国传统艺术品图像及标注数据集

背景与意义：本课题主要研究基于因果图谱的创意问题，需要从数据中归纳创意要素、构建因果图谱，因此建设了中国传统艺术品图像及标注数据集。 内容与来源：数据内容包括100万余幅中国传统艺术品高清图像，部分带相应详细标注。图像的采集方式包括在线爬取、手动下载、截图保存等，原始图像采集后经过清洗、筛选、去噪、裁切、对齐等处理，再进行标注和交叉检查。其中标注内容包括艺术类别、艺术形式、创作年份、展示面积、实际尺寸、作者姓名、作者生卒年份等，标注形式包括现有标注爬取、人工标注等，标注内容受专业人士随机抽检。 产生方法：数据采集地点和标注平台均为线上，采集时间贯穿整个项目执行时间(2020到2022年) 。所需设备为课题研究团队已有基本PC电脑，使用操作系统的基本文件删除操作实现清洗筛选，使用Adobe Photoshop CC实现去噪和对齐，使用Microsoft Excel或WPS表格处理标注数据。 注意事项：本数据集包括了开放获取数据，本数据集共享遵循CC BY-NC-SA（知识共享-署名-非商业性使用-相同方式共享） 4.0 协议。数据仅可用于科学研究活动。

Background and Significance: This project mainly studies creative problems based on causal graphs, which requires inducing creative elements from data and constructing causal graphs. Therefore, a dataset of images and annotations of traditional Chinese artworks has been established.

Content and Source: The dataset contains over 1 million high-definition images of traditional Chinese artworks, some of which are accompanied by detailed corresponding annotations. Image collection methods include online crawling, manual downloading, screenshot saving, etc. After the original images are collected, they undergo processing such as cleaning, screening, denoising, cropping, and alignment, followed by annotation and cross-checking. The annotation content includes art category, art form, creation year, display area, actual size, author's name, author's birth and death years, etc. Annotation methods include existing annotation crawling, manual annotation, etc., and the annotation content is randomly spot-checked by professionals.

Generation Method: The data collection location and annotation platform are both online, and the collection time spans the entire project execution period (2020 to 2022). The required equipment are the basic PCs owned by the research team. Basic file deletion operations of the operating system are used for cleaning and screening, Adobe Photoshop CC is used for denoising and alignment, and Microsoft Excel or WPS Spreadsheets are used to process annotation data.

Notes: This dataset includes open access data, and its sharing follows the CC BY-NC-SA 4.0 (Creative Commons Attribution-NonCommercial-ShareAlike 4.0) license. The data may only be used for scientific research activities.

搜集汇总

背景与挑战

背景概述

该数据集包含100余万幅中国传统艺术品高清图像及专业标注，涵盖多种艺术形式和详细元数据，适用于人工智能和计算机感知研究，遵循CC BY-NC-SA 4.0协议限制商用。