Construction Industry Steel Ordering Lists (CISOL) Dataset

The Construction Industry Steel Ordering Lists (CISOL) dataset comprises table-centric, real-world documents from the construction industry, annotated to facilitate the testing and training of deep learning models for table detection (TD) and table structure recognition (TSR). CISOL Key Features: Steel ordering lists from 24 construction projects carried out between 2015-2023, contributed by 10 distinct German structural engineering firms. Anonymized images to ensure the unrecognizability of specific project or creator information. A total of 3280 images, with 844 annotated following the CISOL annotation guidelines. CISOL is structured into two tracks: Track A: TD-TSR version for end-to-end table detection and table structure recognition tasks. Track B: TSR-only version for table structure recognition tasks, featuring images cropped to the actual table areas with accordingly adjusted annotations. The dataset is developed in accordance with the FAIR Principles, ensuring that it is Findable, Accessible, Interoperable, and Reusable. The CISOL dataset permits expansion following the established annotation guideline. Access to the CISOL Leaderboard will be provided at EvalAI.