RESEARCH · RESEARCH · #2480
Sketch2Sim: end-to-end framework converting hand-drawn circuit photos to SPICE-compatible netlists
The arXiv preprint (arXiv:2610.28836v1) presents Sketch2Sim, a multistage pipeline that converts photographs of hand-drawn resistor–inductor–capacitor (RLC) circuit sketches into executable SPICE-like netlists and computes nodal voltages and currents. The system uses dual YOLO detectors for components and text regions, a CRNN with CTC for text/value recognition, and image processing (thresholding, skeletonization) plus DBSCAN-based clustering to identify nodes; the authors release a curated dataset of 1,000 annotated hand-drawn RLC circuits and report ablation and end-to-end results across varied handwriting and lighting conditions.
KEY POINTS
- The arXiv preprint (arXiv:2610.28836v1) presents Sketch2Sim, a multistage pipeline that converts photographs of hand-drawn resistor–inductor–capacitor (RLC) circuit sketches into executable SPICE-like netlists and computes nodal voltages and currents.
- The system uses dual YOLO detectors for components and text regions, a CRNN with CTC for text/value recognition, and image processing (thresholding, skeletonization) plus DBSCAN-based clustering to identify nodes; the authors release a curated dataset of 1,000 annotated hand-drawn RLC circuits and report ablation and end-to-end results across varied handwriting and lighting conditions.
- This work automates digitizing and simulating hand-drawn circuit diagrams, which can speed education, prototyping, and design review by producing executable netlists directly from photos.
WHY IT MATTERS
This work automates digitizing and simulating hand-drawn circuit diagrams, which can speed education, prototyping, and design review by producing executable netlists directly from photos.