Contour extraction method of SEM image and related product
By extracting coarse and fine contours from multiple images in SEM images and performing two alignments based on the design layout, the problem of insufficient contour extraction accuracy in existing technologies is solved, achieving high-precision contour extraction, meeting the sub-nanometer requirements of OPC models, and supporting accurate modeling of OPC simulation models.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- ORIENTAL CRYSTAL MICROELECTRONICS TECH (SHANGHAI) CO LTD
- Filing Date
- 2026-04-20
- Publication Date
- 2026-07-10
AI Technical Summary
In existing technologies, contour extraction methods for SEM images lack the ability to process complex noise and edge features in images, making it difficult to extract high-precision contours that meet the modeling requirements of OPC models. Especially in semiconductor manufacturing processes below 10nm, the subtle features of lithographic patterns have a significant impact on the electrical performance of chips, and existing methods cannot accurately capture nanoscale edge fluctuations and corner rounding effects.
By acquiring multiple SEM images, coarse and fine contours are extracted separately, and alignment is performed twice based on the design layout. Adaptive binarization and distributed computing are combined to improve the consistency and accuracy of contour extraction. Adaptive threshold binarization algorithms and edge detection algorithms such as Canny and Sobel are used to reduce the impact of electron beam noise and edge blurring.
It achieves high-precision extraction of nanometer-level edge fluctuations and corner rounding effects, meeting the OPC model's requirement for sub-nanometer precision, improving the accuracy and consistency of contour extraction, and supporting accurate modeling of OPC simulation models.
Smart Images

Figure CN122368097A_ABST