Design-Aware Image Search Using Edge Shape Descriptors
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional image search systems are limited in their ability to search for designs or patterns within an input image, leading to poor user experiences as they often return results related to the object instead of the desired design, forcing users to manually sift through numerous images, wasting time and resources.
Innovation Solution
The design-aware image search system processes input images to extract and isolate designs or patterns, generating shape descriptors that are compared to other images to locate similar or matching designs, allowing for the identification of products that incorporate these designs without returning irrelevant object results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional image search systems use all objects in an input image for search, then the search process is simple, but the search results are inaccurate and return irrelevant objects instead of the desired design
Solution Approach 1:
The patent segments the input image into multiple components: the target object, the design/pattern on the object, and the background. By separating the design extraction from the object itself, the system can search for designs independently of the object type, thereby improving search result accuracy while managing complexity through modular processing steps
Solution Approach 2:
The patent extracts the design or pattern from the object in the input image by creating an edge image and generating shape descriptors that represent only the design features. This extraction process removes the object identity from the search query, allowing the system to find similar designs across different object types, thus improving accuracy without requiring complex object recognition
2Loss of time
If users manually sift through numerous search results to find desired designs, then the search system can use simple object-based matching, but user time and interaction resources are wasted
Solution Approach 1:
The patent performs preliminary design extraction and shape descriptor generation before the search process. By pre-processing the input image to isolate and characterize the design features, the system prepares the search query in advance with precise design-specific descriptors, enabling accurate matching without requiring users to manually filter through irrelevant results
Solution Approach 2:
The patent replaces the manual mechanical process of sifting through images with an automated image processing system that uses edge detection and shape descriptor comparison. This substitution automates the design matching process, significantly reducing user time and interaction while maintaining high precision through algorithmic comparison of shape features
3Adaptability or versatility
If conventional systems focus search on the entire object, then the processing is straightforward, but the system cannot locate designs on different types of objects
Solution Approach 1:
The patent changes the search parameter from object-based to design-based by transforming the input image into shape descriptors that represent geometric features of the design. This parameter transformation allows the system to search for designs across different object types and categories, greatly improving versatility while managing extraction difficulty through standardized edge detection and descriptor generation processes
Data Source
AI summary
Systems and techniques for a design-aware image search are described. The design-aware image search techniques described herein capture a design on an item to determine additional items with similar or matching designs. An image of the item is used to create an edge image of a design, and shape descriptors are generated describing features of the edge image. These shape descriptors are compared to shape descriptors associated with other images to locate images that have similar or matching designs as compared with the input image. The design-aware image search system may use these relationships to generate a search result with images or products having a design similar to the design on the input image.