J. Chen, Yibo Lin, Yufeng Guo, Maolin Zhang, M. Alawieh, D. Pan
{"title":"Lithography hotspot detection using a double inception module architecture","authors":"J. Chen, Yibo Lin, Yufeng Guo, Maolin Zhang, M. Alawieh, D. Pan","doi":"10.1117/1.JMM.18.1.013507","DOIUrl":null,"url":null,"abstract":"Abstract. With the shrinking feature sizes of semiconductor devices, manufacturing challenges increase dramatically. Among these challenges, lithography hotspot stands out as a prominent ramification of the growing gap between design and manufacturing. Practically, a hotspot refers to the failure in printing desired patterns in lithography. As lithography hotspots have significant impacts on manufacturing yield, the detection of hotspots in the early design stage is desired to achieve fast design closure. We propose a lithography hotspot detection framework using a double inception module structure. This structure performs better in both accuracy and false alarms by widening the conventional stacked structure to benefit feature extraction and using global average pooling to keep the spatial information. Experimental results show that the proposed structure achieves better performance than existing methods.","PeriodicalId":16522,"journal":{"name":"Journal of Micro/Nanolithography, MEMS, and MOEMS","volume":"124 1","pages":"013507 - 013507"},"PeriodicalIF":1.5000,"publicationDate":"2019-03-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"11","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"Journal of Micro/Nanolithography, MEMS, and MOEMS","FirstCategoryId":"101","ListUrlMain":"https://doi.org/10.1117/1.JMM.18.1.013507","RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q3","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 11
Abstract
Abstract. With the shrinking feature sizes of semiconductor devices, manufacturing challenges increase dramatically. Among these challenges, lithography hotspot stands out as a prominent ramification of the growing gap between design and manufacturing. Practically, a hotspot refers to the failure in printing desired patterns in lithography. As lithography hotspots have significant impacts on manufacturing yield, the detection of hotspots in the early design stage is desired to achieve fast design closure. We propose a lithography hotspot detection framework using a double inception module structure. This structure performs better in both accuracy and false alarms by widening the conventional stacked structure to benefit feature extraction and using global average pooling to keep the spatial information. Experimental results show that the proposed structure achieves better performance than existing methods.