Hugging Face's logo

Datasets:
danqing-ai
/
NPM-Artifacts-zh

Languages:
Chinese
English
Tags:
art
image-captioning
text-to-image
cultural-heritage
DOI:
License:
Dataset card Files Community
2
Dataset Viewer
The dataset viewer is not available for this subset.
Cannot get the split names for the config 'default' of the dataset.
Exception:    SplitsNotFoundError
Message:      The split names could not be parsed from the dataset config.
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/inspect.py", line 289, in get_dataset_config_info
                  for split_generator in builder._split_generators(
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/packaged_modules/folder_based_builder/folder_based_builder.py", line 237, in _split_generators
                  raise ValueError(
              ValueError: `file_name` or `*_file_name` must be present as dictionary key (with type string) in metadata files
              
              The above exception was the direct cause of the following exception:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/split_names.py", line 65, in compute_split_names_from_streaming_response
                  for split in get_dataset_split_names(
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/inspect.py", line 343, in get_dataset_split_names
                  info = get_dataset_config_info(
                File "/src/services/worker/.venv/lib/python3.9/site-packages/datasets/inspect.py", line 294, in get_dataset_config_info
                  raise SplitsNotFoundError("The split names could not be parsed from the dataset config.") from err
              datasets.inspect.SplitsNotFoundError: The split names could not be parsed from the dataset config.

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

NPM-Artifacts-zh: National Palace Museum Open Artifacts Dataset

Dataset Description

This project collects and organizes public artifact data from the National Palace Museum Open Data Platform. The dataset contains high-resolution images of artifacts and their corresponding rich, structured metadata. All metadata is in Traditional Chinese, detailing information such as the artifact's name, dynasty, dimensions, materials, inscriptions, and seal impressions.

This dataset aims to provide researchers, developers, and cultural enthusiasts with a valuable and accessible resource for research and application in various fields, including computer vision, natural language processing, and digital humanities.

Dataset Content

The dataset primarily consists of the following parts:

  1. metadata.csv: A CSV file containing the master table of all artifact metadata. Each row corresponds to one artifact, with fields serialized as JSON strings.

  2. images/: A directory containing all artifact images, including main views and detailed shots.

Data Field Descriptions (metadata.csv & data/*.json)

  • UniqueID: A unique ID for the artifact.
  • URL: The original URL of the artifact's detail page.
  • 文物名称 (Artifact Name): The official name of the artifact.
  • 基本資料 (Basic Information): A JSON object containing basic info like "item name" (品名), "artist" (作者), "dynasty" (朝代), and "format/dimensions" (格式/尺寸).
  • 典藏尺寸 (Collection Dimensions): A JSON array containing detailed dimension information such as "form" (形式), "painting dimensions" (画幅尺寸), and "mounting dimensions" (裱绫尺寸).
  • 质地 (Material): A JSON array containing "primary material" (主要质地) and "secondary material" (次要质地).
  • 题跋资料 (Inscription Data): A JSON array containing information about inscriptions, including the inscriber, content, and source.
  • 印记资料 (Seal Data): A JSON array containing information about seals, including the seal script transcription and source.
  • 主题 (Subjects): A JSON array of keywords describing the artifact's subject matter (e.g., landscape, flowers and birds, figures).
  • 技法 (Techniques): A JSON array of keywords describing the techniques used (e.g., color, ink wash, gold outline).
  • 參考資料 (References): A list of bibliographic references.
  • 保存維護 (Conservation and Maintenance): Information about the artifact's conservation and maintenance.

Uses and Tasks

This dataset can be used for a variety of tasks, including but not limited to:

  • Image Classification: Classifying artifacts by their dynasty, artist, or subject based on their images.
  • Image Captioning: Training models to generate descriptive text for artifact images.
  • Text-to-Image Generation: Using descriptive text from the metadata to generate corresponding artifact images.
  • Optical Character Recognition (OCR): Recognizing text from inscriptions and seals on paintings and calligraphy.
  • Digital Humanities Research: Conducting quantitative analysis and visualization of various artifact metadata.

Authors

License and Citation

The content of this dataset is sourced from the National Palace Museum Open Data Platform and is governed by the "Open Government Data License, version 1.0 (政府資料開放授權條款1.0版)".

Under this license, you are free to reproduce, distribute, publicly transmit, publicly present, publish, utilize, or create derivative works from this dataset without fees or separate application.

When using the data, please note:

  1. You must not maliciously alter or distort the data content.
  2. You should preserve the attribution of the author's name (if applicable).
  3. It is recommended to credit the data source as the "National Palace Museum" upon use.

How to Cite: If you use this dataset in your research, we recommend citing it as follows:

The dataset used in this research was collected and organized by the "NPM-Artifacts-zh" project from the National Palace Museum Open Data Platform (https://digitalarchive.npm.gov.tw/opendata/).

NPM-Artifacts-zh: 台北故宫博物院开放文物数据集

数据集描述

本项目收集并整理了来自台北故宫博物院开放数据平台的公开文物数据。数据集包含高解析度的文物图像以及与之对应的、丰富的结构化元数据。元数据均为繁体中文，详细描述了文物的名称、年代、尺寸、材质、题跋、印记等信息。

这个数据集旨在为研究者、开发者和文化爱好者提供一份易于访问和使用的宝贵资源，可用于计算机视觉、自然语言处理、数字人文等多个领域的研究与应用。

数据集内容

数据集主要包含以下几个部分：

  1. metadata.csv: 包含所有文物元数据总表的CSV文件。每一行对应一件文物，字段被序列化为JSON字符串。

  2. images/: 存放所有文物图片的目录，包含主图和细节图。

数据字段说明 (metadata.csv & data/*.json)

  • UniqueID: 文物唯一ID。
  • URL: 原始的文物详情页链接。
  • 文物名称: 文物的官方名称。
  • 基本資料: 包含“品名”、“作者”、“朝代”、“格式/尺寸”等基本信息的JSON对象。
  • 典藏尺寸: 包含“形式”、“画幅尺寸”、“裱绫尺寸”等详细尺寸信息的JSON数组。
  • 质地: 包含“主要质地”、“次要质地”的JSON数组。
  • 题跋资料: 包含题跋者、内容、资料来源等的JSON数组。
  • 印记资料: 包含印记释文、资料来源等的JSON数组。
  • 主题: 描述文物主题的关键词JSON数组 (例如：山水、花鸟、人物等)。
  • 技法: 描述文物所用技法的关键词JSON数组 (例如：设色、水墨、描金等)。
  • 參考資料: 参考文献列表。
  • 保存維護: 文物的保存与维护信息。

用途与任务

本数据集可用于多种任务，包括但不限于：

  • 图像分类: 根据文物图片对其朝代、作者或主题进行分类。
  • 图像描述生成 (Image Captioning): 训练模型为文物图片生成描述性文字。
  • 文本到图像生成 (Text-to-Image): 使用元数据中的描述文字来生成对应的文物图像。
  • 光学字符识别 (OCR): 对书画作品中的题跋、印记进行文字识别。
  • 数字人文研究: 对文物的各种元数据进行量化分析和可视化。

作者

授权与引用

本数据集的内容来源于台北故宫博物院开放数据平台，遵循“政府資料開放授權條款1.0版 (Open Government Data License, version 1.0)”。

根据该条款，您可以自由地重制、散布、公开传输、公开发表、发行、利用本数据集，或进行各种衍伸创作，无需付费、也无需另外申请。

使用时请注意:

  1. 不得恶意变更、扭曲数据内容。
  2. 应保留作者的姓名表示权（如适用）。
  3. 建议在使用时标注数据来源为“台北故宫博物院”。

如何引用: 如果您在研究中使用了本数据集，建议按以下方式引用：

本研究使用的数据集由 "NPM-Artifacts-zh" 项目收集、整理自台北故宫博物院开放数据平台 (https://digitalarchive.npm.gov.tw/opendata/)。
Downloads last month
1,250