Pattern Similarity Metrics for Layout Pattern Classification and Their Validity Analysis by Lithographic Responses

In order to detect manufacturing hotspots from huge layout pattern efficiently or to characterize hotspots effectively, various layout pattern classification methods had been proposed. Layout pattern clustering using layout pattern similarity metrics such as area-difference-ratio and edge-shifting-distance could be used as preprocessing to shorten the runtime of classification significantly. However the validity of classification will be degraded if the parameters used for similarity metrics are not appropriately chosen. In this paper, the layout pattern classification methods and layout pattern similarity metrics proposed so far are summarized and are discussed.