# Modular Graph Extraction for Handwritten Circuit Diagram Images

arXiv: https://arxiv.org/abs/2402.11093 · HTML: https://mlsanity.com/paper/2402.11093/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 6 of its 10 references that are in the map.
- First posted on arXiv 16 Feb 2024.

## Builds on (most cited first, 6 of 6)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [Hand-Drawn Electrical Circuit Recognition using Object Detection and Node Recognition](https://mlsanity.com/paper/2106.11559/) (2021)
- [A Public Ground-Truth Dataset for Handwritten Circuit Diagram Images](https://mlsanity.com/paper/2107.10373/) (2021)
- [Instance Segmentation Based Graph Extraction for Handwritten Circuit Diagram Images](https://mlsanity.com/paper/2301.03155/) (2023)

## Built on by (most cited first, 1 of 1)

- [Dataset Construction for Training LLM to Learn Analog Circuit Knowledge](https://mlsanity.com/paper/2508.10409/) (2025)

## Abstract (arXiv)

As digitization in engineering progressed, circuit diagrams (also referred to as schematics) are typically developed and maintained in computer-aided engineering (CAE) systems, thus allowing for automated verification, simulation and further processing in downstream engineering steps. However, apart from printed legacy schematics, hand-drawn circuit diagrams are still used today in the educational domain, where they serve as an easily accessible mean for trainees and students to learn drawing this type of diagrams. Furthermore, hand-drawn schematics are typically used in examinations due to legal constraints. In order to harness the capabilities of digital circuit representations, automated means for extracting the electrical graph from raster graphics are required. While respective approaches have been proposed in literature, they are typically conducted on small or non-disclosed datasets. This paper describes a modular end-to-end solution on a larger, public dataset, in which approaches for the individual sub-tasks are evaluated to form a new baseline. These sub-tasks include object detection (for electrical symbols and texts), binary segmentation (drafter's stroke vs. background), handwritten character recognition and orientation regression for electrical symbols and texts. Furthermore, computer-vision graph assembly and rectification algorithms are presented. All methods are integrated in a publicly available prototype.