I looked into this last year. The best you can do is import the schematic to a CAD app and trace over it. The main issue is even though we can see the symbols for passives and parts and interpret it correctly with our eyes. Not all schematics use the same style/size/shape symbols. So you’ll need machine learning on top of OCR. I think this would be an awesome thesis project or something for the more computer vision savvy on here.<div><br></div><div>Want I want is a app that you can use your camera phone pointed at a schematic and it will create augmented reality and show the signal flow and voltage directions, etc. </div><div><br></div><div>Instead we get stupid AR face jewelry and morphs. lame</div><div><br></div><div>Kylee Kennedy<br><br>On Monday, November 2, 2020, Jimmy Moore <<a href="mailto:jamoore84@gmail.com">jamoore84@gmail.com</a>> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr"><div>Does anyone know whether any tools or techniques exist for converting a physical circuit sketch to a digital representation? I'm imagining something like OCR, but for circuit symbols?<br></div><div><br></div><div>I have next to zero background in this area, but was still surprised by how little work exists in this area:</div><div><br><div><span style="text-align:left" dir="ltr"> <a rel="nofollow noreferrer noopener" href="https://www.google.com/url?q=https://hcommons.org/deposits/download/hc:29916/CONTENT/turkel-sirohey_csdh_2020.pdf/&sa=D&source=hangouts&ust=1604468546636000&usg=AFQjCNFQOw2MbgKjUBmQLmoLvTvD4N7Zsg" dir="ltr" target="_blank">https://hcommons.org/deposits/<wbr>download/hc:29916/CONTENT/<wbr>turkel-sirohey_csdh_2020.pdf/</a><br></span></div><div><span style="text-align:left" dir="ltr"><br></span></div><div><span style="text-align:left" dir="ltr">My own google-scholaring returns just a handful of papers, none of which have many citations. I understand that this is a bit of an atypical use-case in the professional world, but I would have expected more, especially with machine learning advances.<br></span></div><div><span style="text-align:left" dir="ltr"><br></span></div><div><span style="text-align:left" dir="ltr">Am I looking in the wrong place? Is there anyone on this list who can authoritatively say this is much more complex than I understand?<br></span></div></div></div>
</blockquote></div>