Feature-Based Image Detection for Visual IP Infringement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing visual search systems for identifying objects subject to visual intellectual property rely heavily on text-based keywords and lack graphical user interfaces and computer vision techniques to focus on user-selected features, making it difficult to efficiently identify potential infringement in electronic catalogs.
Innovation Solution
The development of improved graphical user interfaces and computer vision systems that allow users to search images based on user-selected features, using convolutional neural networks to extract feature vectors and perform nearest neighbor searches, enabling efficient identification of similar images with a focus on user-defined coordinates and features.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If text-based keywords are used for visual search, then the search system is simpler to implement, but the ability to focus on user-selected features is lost
Solution Approach 1:
The patent replaces text-based keyword search with computer vision technology. Users can directly interact with images by selecting features of interest, and the system uses convolutional neural networks to automatically extract and compare visual features, eliminating the need for text input and providing more intuitive visual search capability
Solution Approach 2:
The patent introduces a graphical user interface as an intermediary between the user and the search system. The GUI allows users to visually select features on images while the system processes these selections through computer vision algorithms, creating a bridge between user intent and automated feature extraction
2Productivity
If existing visual search systems are used, then the system can search images, but it cannot efficiently identify potential infringement in electronic catalogs
Solution Approach 1:
The patent enables users to select specific local features or regions of interest within images. Instead of searching based on overall image appearance, the system allows focused search on particular features such as logos, patterns, or design elements, improving both the precision of infringement detection and the efficiency of reviewing potential matches
Solution Approach 2:
The patent changes the search parameters from text-based keywords to visual feature vectors extracted by convolutional neural networks. The system compares these feature vectors to identify similar images, enabling more accurate and efficient detection of visual intellectual property infringements in electronic catalogs
Data Source
AI summary
System and methods are provided for improved visual search systems that can use local features of images. The search system provides a user interface that enables a user to select areas or portions of an image to search on the selected features and the overall appearance of the image. The search system further provides customized user interfaces to exclude certain portions of images from the search algorithms. The search system can be used to detect potential intellectual property risks associated with items in an electronic catalog.