Recent developments in LSI mask data processing show a tendency towards unpreceded data volume increases. This paper introduces a data compression technique designed to reduce mask data volumes and is based on geometrical library recognition in which a library is defined as a set of patterns that can be referenced as a single object repeated along a 2 dimensional mesh. The data compression technique is referred to as mask data geometrical library recognition compression (MD-GLRC) and includes a newly developed repetition center tracking (RCT) and library inheritance tree (LIT) algorithm. By applying MD-GLRC, a 885 Mbyte 0.25 μm memory device mask data sample was compressed to 7 Mbytes in 3.5 CPU hours on a 135 MIPS workstation.
"Geometrical library recognition for mask data compression", Proc. SPIE 2793, Photomask and X-Ray Mask Technology III, (24 July 1996); doi: 10.1117/12.245219; https://doi.org/10.1117/12.245219