?url_ver=Z39.88-2004&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Adc&rft.title=CADTalk%3A+An+Algorithm+and+Benchmark+for+Semantic+Commenting+of+CAD+Programs&rft.creator=Yuan%2C+H&rft.creator=Xu%2C+J&rft.creator=Pan%2C+H&rft.creator=Bousseau%2C+A&rft.creator=Mitra%2C+NJ&rft.creator=Li%2C+C&rft.description=CAD+programs+are+a+popular+way+to+compactly+encode+shapes+as+a+sequence+of+operations+that+are+easy+to+para-metrically+modify.+However%2C+without+sufficient+semantic+comments+and+structure%2C+such+programs+can+be+challenging+to+understand%2C+let+alone+modify.+We+introduce+the+problem+of+semantic+commenting+CAD+programs%2C+wherein+the+goal+is+to+segment+the+input+program+into+code+blocks+corresponding+to+semantically+meaningful+shape+parts+and+assign+a+semantic+label+to+each+block.+We+solve+the+problem+by+combining+program+parsing+with+visual-semantic+analysis+afforded+by+recent+advances+in+foundational+language+and+vision+models.+Specifically%2C+by+executing+the+input+programs%2C+we+create+shapes%2C+which+we+use+to+generate+conditional+photorealistic+images+to+make+use+of+semantic+annotators+for+such+images.+We+then+distill+the+information+across+the+images+and+link+back+to+the+original+programs+to+semantically+comment+on+them.+Additionally%2C+we+collected+and+annotated+a+benchmark+dataset%2C+CADTalk%2C+consisting+of+5%2C288+machine-made+programs+and+45+human-made+programs+with+ground+truth+semantic+comments.+We+exten-sively+evaluated+our+approach%2C+compared+it+to+a+GPT-based+baseline%2C+and+an+open-set+shape+segmentation+baseline%2C+and+reported+an+83.24%25+accuracy+on+the+new+CADTalk+dataset.+Code+and+data%3A+https%3A%2F%2Fenigma-li.github.io%2FCADTalk%2F.&rft.publisher=IEEE&rft.date=2024-09-16&rft.type=Proceedings+paper&rft.language=eng&rft.source=+++++In%3A++Proceedings+of+the+IEEE+Computer+Society+Conference+on+Computer+Vision+and+Pattern+Recognition.++(pp.+pp.+3753-3762).++IEEE%3A+Seattle%2C+WA%2C+USA.+(2024)+++++&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10199675%2F1%2FYuan_CADTalk_An_Algorithm_and_Benchmark_for_Semantic_Commenting_of_CAD_CVPR_2024_paper.pdf&rft.format=text&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10199675%2F7%2FYuan_CADTalk_An_Algorithm_CVPR_2024_supplemental.pdf&rft.identifier=https%3A%2F%2Fdiscovery.ucl.ac.uk%2Fid%2Feprint%2F10199675%2F&rft.rights=open