PCBnet: A Dataset and Automatic Construction of SPICE Netlists from Schematic Images
Read the original on arXiv Computer Vision →PCBnet is a new large‑scale dataset of printed circuit board (PCB) schematics that includes over 300 real‑world designs, more than 50,000 component instances, 150,000 wires, 100,000 text regions, and 400,000 characters, each paired with a SPICE netlist. The authors also introduce an automated pipeline that converts schematic images into netlists, achieving 94.54% component detection mAP, 98.57% text recognition accuracy, and 84.47% end‑to‑end connectivity accuracy. This resource aims to serve as a benchmark and data foundation for future AI‑driven PCB design automation.
Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computer Vision.