An annotated synthetic dataset of 500 Piping and Instrumentation Diagrams (P&IDs) containing complex symbols and noise. The dataset is designed for the object detection task, focusing on symbols only. It was created by Paliwal et al. and published in 2021.
Use Cases
- Training object detection models to identify symbols in P&IDs based on the described synthetic data.
- Benchmarking model robustness against different types of noise incorporated into the diagrams.
- Developing automated tools for digitizing industrial piping and instrumentation documentation.
Strengths
- 500 annotated diagrams provide a defined scale for model training.
- Incorporates different types of noise and complex symbols, as described, to test model robustness.
- Dataset is specifically annotated for the object detection task.
Limitations
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
- Data may reflect bias inherent to the synthetic generation method.
Provenance
- Source
- Paliwal, S., Jain, A., Sharma, M., & Vig, L. (2021). Digitize-PID: Automatic Digitization of Piping and Instrumentation Diagrams. ArXiv, abs/2109.03794.
- Collection Method
- Synthetic dataset generation.
- Time Range
- Publication date is 2021.
- Freshness
- Last updated 2026-01-23 10:05:31; freshness should be verified.
- Geography
- null