Curvilinear Layout Clustering via Density Feature Vectors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional layout pattern classification tools struggle with accuracy and performance when processing curvilinear shapes, which are increasingly used in silicon photonics and memory chip design, due to their reliance on Manhattan shape processing methods.
Innovation Solution
A method involving the extraction of density feature vectors using a set of non-overlapping rings, followed by a machine learning-based clustering process using hyperboxes in a hyperspace to separate and subcluster layout features, with properties such as curvature, linewidths, and spacing, and incorporating lithography simulation for error detection and correction.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional layout pattern classification tools are used to process curvilinear shapes, then the tools can handle Manhattan shapes effectively, but the accuracy and performance deteriorate when processing curvilinear shapes
Solution Approach 1:
The patent transforms the layout pattern classification problem by changing the parameter representation from traditional Manhattan-based metrics to density feature vectors that capture curvilinear characteristics. Density feature vectors compute the distribution of layout features within concentric rings, enabling accurate representation of curved shapes while maintaining computational efficiency for high-dimensional data classification
Solution Approach 2:
The patent extends the classification approach by mapping density feature vectors into a hyperspace using machine learning techniques. This dimensional transformation allows the system to handle the increased complexity of curvilinear patterns by operating in a higher-dimensional feature space where curvilinear shapes can be effectively separated and classified
2Measurement precision
If machine learning-based clustering with density feature vectors is implemented, then curvilinear features are handled accurately, but the computational complexity increases
Solution Approach 1:
The patent segments the computational problem by dividing the layout into multiple concentric rings and computing density features independently for each ring. This segmentation transforms a complex global classification problem into multiple simpler local density calculations, reducing overall computational complexity while preserving curvilinear feature accuracy
Solution Approach 2:
The patent uses density feature vectors as a simplified copy or representation of the actual layout patterns. Instead of directly processing complex curvilinear geometries, the system creates compact density-based representations that capture essential features, enabling efficient machine learning classification without requiring complex geometric computations
Data Source
AI summary
Various aspects of the present disclosed technology relate to techniques for classifying layout patterns. First, a set of density feature vectors for a set of layout regions in the layout design are extracted using a set of rings. Each component of a density feature vector in the set of density feature vectors corresponds to a ring in the set of rings. The set of rings do not overlap with each other and cover a whole area of a circle when being placed together. Next, a machine learning-based clustering process is performed to separate layout features in the set of layout regions into clusters of layout features based on the set of density feature vectors. Each of the clusters of layout features may be further divided into subclusters based on one or more properties.