Design Image Clustering Using Similarity-Based Search Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face difficulty in efficiently searching for design images due to the large number of reference images stored in databases, and existing search methods like BIG framework struggle to accurately identify relevant images based on design elements.
Innovation Solution
A design image clustering method that groups reference images into clusters based on similarity, using a processor to calculate and display representative images and similarity metrics for each cluster, facilitating effective search support.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional image search methods are used to search through a large database of design images, then the search process becomes time-consuming and inefficient, but implementing advanced clustering algorithms increases system complexity
Solution Approach 1:
The patent segments the large database of design images into multiple clusters based on visual similarity. Each cluster contains images with comparable characteristics, allowing users to navigate through organized groups rather than searching the entire database sequentially. This segmentation significantly improves search efficiency by reducing the search space while maintaining system manageability through modular cluster structures.
2Ease of operation
If design images are clustered into multiple groups based on similarity, then user search efficiency improves, but the computational resources and time required for clustering increase
Solution Approach 1:
The patent performs clustering computation in advance before users perform searches. The system pre-processes the design image database, calculates similarity metrics, and organizes images into clusters beforehand. This preliminary action transforms the computationally intensive clustering task into a one-time operation, making subsequent user searches much faster and more efficient without requiring real-time computation during actual search operations.
3Measurement precision
If detailed similarity calculations are performed between target design element information and reference design images, then clustering accuracy improves, but the processing time and computational load increase
Solution Approach 1:
The patent applies local quality by calculating similarity metrics selectively for specific design elements within images rather than analyzing entire images uniformly. The system identifies and compares key design elements (such as shapes, colors, patterns) that are most relevant for clustering, assigning different weights to different elements based on their importance. This approach maintains high clustering accuracy for critical features while reducing overall computational load by focusing processing power on locally significant characteristics.
Data Source
AI summary
A design image clustering method comprises the steps of: calculating the degree of similarity between pieces of target design element information selected by a user among pieces of design element information about a target design image and a plurality of reference design images stored in a database; clustering, on the basis of the degree of similarity, the reference design images into a plurality of clusters; and generating, for each cluster, at least one of at least one representative design image, an overlap image for the representative design image, and the degree of similarity between a center vector of each cluster and the target design image.