SynthPID: P&ID digitization from Topology-Preserving Synthetic Data

TL;DR

SynthPID introduces a large synthetic dataset of P&IDs with real topology, enabling effective digitization of diagrams without real data during training.

Contribution

The paper presents SynthPID, a synthetic P&ID dataset with topology seeded from real drawings, and a model that achieves high accuracy without real training data.

Findings

Synthetic data with real topology improves edge detection accuracy.

Model trained on SynthPID achieves 63.8% edge mAP without real data.

Performance plateaus beyond 400 synthetic images, highlighting seed diversity importance.

Abstract

Automating the digitization of Piping and Instrumentation Diagrams (P&IDs) into structured process graphs would unlock significant value in plant operations, yet progress is bottlenecked by a fundamental data problem: engineering drawings are proprietary, and the entire community shares a single public benchmark of just 12 annotated images. Prior attempts at synthetic augmentation have fallen short because template-based generators scatter symbols at random, producing graphs that bear little resemblance to real process plants and, accordingly, yield only approximately 33% edge detection accuracy under synth-only training. We argue the failure is structural rather than visual and address it by introducing SynthPID, a corpus of 665 synthetic P&IDs whose pipe topology is seeded directly from real drawings. Paired with a patch-based Relationformer adapted for high-resolution diagrams, a model trained on SynthPID alone achieves 63.8 +/- 3.1% edge mAP on PID2Graph OPEN100 without seeing a single real P&ID during training, closing within 8 pp of the real-data oracle. These gains hold up under a controlled comparison against the template-based regime, confirming that generation quality drives performance rather than model choice. A scaling study reveals that gains flatten beyond roughly 400 synthetic images, pointing to seed diversity as the binding constraint.