Invariant Clustering of Circuit Images for EDA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional unsupervised learning techniques and clustering algorithms fail to accurately group circuit images that differ only by accidental or unintentional rotations or reflections, leading to reduced accuracy in electronic design automation (EDA) analyses.
Innovation Solution
The implementation of an image clustering engine that uses unsupervised learning with an invariant property set to ignore specific image transformations such as rotations and reflections, allowing for invariant clustering of circuit images, which increases the accuracy of pattern classifications and EDA analyses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional unsupervised learning techniques and clustering algorithms are used to group circuit images, then the clustering process is simple to implement, but the accuracy of grouping is reduced due to accidental rotations or reflections
Solution Approach 1:
The patent segments the clustering process into two distinct stages: (1) feature extraction that is invariant to rotation and reflection transformations, and (2) clustering based on these invariant features. This segmentation allows the system to maintain high clustering accuracy by ignoring accidental transformations while keeping the overall system manageable through modular design.
Solution Approach 2:
The patent applies preliminary action by pre-processing circuit images to extract rotation- and reflection-invariant features before performing clustering. By performing this invariant feature extraction in advance, the system eliminates the negative impact of accidental transformations before the clustering algorithm is applied, thereby improving accuracy without requiring the clustering algorithm itself to be complex.
2Reliability
If rotation- and reflection-invariant clustering is implemented, then the accuracy of EDA analyses is improved, but the computational complexity increases
Solution Approach 1:
The patent extracts only the essential invariant features (rotation- and reflection-invariant) from circuit images, taking out the relevant information needed for reliable EDA analysis while discarding irrelevant transformation details. This extraction approach improves reliability by focusing on meaningful patterns while avoiding the computational burden of processing all possible image variations.
Solution Approach 2:
The patent changes the parameter representation of circuit images by transforming them into rotation- and reflection-invariant feature spaces. This parameter transformation allows the system to achieve higher EDA analysis reliability by comparing circuits based on their intrinsic properties rather than their specific orientations or reflections, while the dimensionality reduction in feature space helps manage computational complexity.
3Measurement precision
If conventional clustering algorithms are used without invariant properties, then the processing speed is fast, but the accuracy of pattern classifications is reduced
Solution Approach 1:
The patent performs preliminary action by pre-computing rotation- and reflection-invariant features from circuit images before the clustering and pattern classification steps. This pre-processing ensures that when clustering is performed, the algorithm already operates on transformed features that eliminate the impact of accidental transformations, thereby improving pattern classification accuracy without requiring repeated transformations during the clustering process itself.
Solution Approach 2:
The patent transitions from operating in the original image space to operating in a transformed feature space that is invariant to rotation and reflection. This dimensionality change allows the system to achieve higher pattern classification accuracy by comparing circuits in a space where accidental transformations do not create artificial differences, while the reduced dimensionality of the feature space helps maintain reasonable processing speeds.
Data Source
AI summary
A system may include an image clustering engine and a cluster provision engine. The image clustering image may be configured to access a set of circuit images and cluster the circuit images into different groups via an unsupervised learning process, wherein clustering by the unsupervised learning process is invariant to each invariant property of an invariant property set. A given invariant property in the invariant property set may correspond to a given image transformation, the invariant properties in the invariant property set may be discrete, and the total number of invariant properties in the invariant property set may be finite. The cluster provision engine may be configured to provide the clustered circuit images for further processing or analysis by an electronic design automation (EDA) application.