An improved ridge extraction algorithm for the development of lithographic masks

Skeleton extraction, also referred as ridge extraction, plays an important role in various semiconductor industry applications, particularly within Electronic Design Automation (EDA). It is crucial for accurately representing the essential structure of complex/curvilinear layouts used in manufacturing. However, traditional skeletonization algorithms often encounter issues with maintaining both skeleton connectivity and accuracy, especially when extracted using a high precision distance map. These challenges can result in disconnected or distorted skeletons, which compromise the efficiency and precision of downstream processes. This paper introduces an improved skeleton extraction algorithm designed to overcome these limitations. Built upon a gradient-based approach, we have incorporated additional techniques and rules to improve the algorithm’s ability to handle noise and ensure full skeleton connectivity. By addressing the common issues found in earlier methods, the proposed algorithm ensures that the extracted skeleton is both correct and continuous. This improvement guarantees a more reliable representation of layout structures. In the semiconductor industry, skeleton extraction plays a vital role in critical processes such as Optical Proximity Correction (OPC) and mask correction. These processes require precise, connected skeletons to optimize the photomask patterns used in lithography, ensuring the accurate transfer of designs onto silicon wafers. By enhancing the robustness and accuracy of skeleton extraction, our algorithm contributes to higher precision and improved efficiency in semiconductor manufacturing, leading to better pattern fidelity and increased yield.