Teklia/NewsEye-Austrian-line

--- license: mit language: - de task_categories: - image-to-text pretty_name: NewsEye-Austrian-line dataset_info: features: - name: image dtype: image - name: text dtype: string splits: - name: train num_examples: 51588 - name: validation num_examples: 4379 dataset_size: 55967 tags: - atr - htr - ocr - historical - printed --- # NewsEye Austrian - line level ## Table of Contents - [NewsEye Austrian - line level](#newseye-austrian-line-level) - [Table of Contents](#table-of-contents) - [Dataset Description](#dataset-description) - [Languages](#languages) - [Dataset Structure](#dataset-structure) - [Data Instances](#data-instances) - [Data Fields](#data-fields) ## Dataset Description - **Homepage:** [NewsEye project](https://www.newseye.eu/) - **Source:** [Zenodo](https://zenodo.org/records/3387369) - **Point of Contact:** [TEKLIA](https://teklia.com) ## Dataset Summary The dataset comprises Austrian newspaper pages from 19th and early 20th century. The images were provided by the Austrian National Library. ### Languages The documents are in Austrian German with the Fraktur font. Note that all images are resized to a fixed height of 128 pixels. ## Dataset Structure ### Data Instances ``` { 'image': <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=4300x128 at 0x1A800E8E190, 'text': 'Mann; und als wir uns zum Angriff stark genug' } ``` ### Data Fields - `image`: a PIL.Image.Image object containing the image. Note that when accessing the image column (using dataset[0]["image"]), the image file is automatically decoded. Decoding of a large number of image files might take a significant amount of time. Thus it is important to first query the sample index before the "image" column, i.e. dataset[0]["image"] should always be preferred over dataset["image"][0]. - `text`: the label transcription of the image.

原始信息汇总

NewsEye Austrian - line level 数据集概述

数据集描述

该数据集包含19世纪和20世纪初的奥地利报纸页面，图像由奥地利国家图书馆提供。文档语言为奥地利德语，使用Fraktur字体。所有图像被调整为固定高度128像素。

数据集结构

数据实例

json { "image": <PIL.JpegImagePlugin.JpegImageFile image mode=RGB size=4300x128 at 0x1A800E8E190>, "text": "Mann; und als wir uns zum Angriff stark genug" }

数据字段

• image: 包含图像的PIL.Image.Image对象。访问图像列时（例如 dataset[0]["image"]），图像文件会自动解码。解码大量图像文件可能需要较长时间，因此建议先查询样本索引再访问"image"列，即 dataset[0]["image"] 应优先于 dataset["image"][0]。
• text: 图像的标签转录文本。

数据集信息

• 特征:
  • image: 图像数据类型
  • text: 字符串数据类型
• 分割:
  • train: 51588个样本
  • validation: 4379个样本
• 数据集大小: 55967个样本
• 标签:
  • atr
  • htr
  • ocr
  • historical
  • printed