Image Query System Using Visual Similarity Indexing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image search technologies rely on text queries, which can be cumbersome for users who struggle to describe images or prefer not to use text inputs, especially on mobile devices, and lack efficient methods for retrieving visually similar images.
Innovation Solution
The method involves generating a searchable index of images using local descriptors, building a distributed kd-tree infrastructure, and calculating n-grams and their scores to derive a textual query from an image, allowing for the retrieval of visually similar images and presenting relevant search results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If text queries are used for image search, then search functionality is provided, but user convenience deteriorates especially on mobile devices
Solution Approach 1:
The patent replaces the mechanical text input system with an image-based query system. Users can upload images directly instead of typing text descriptions, substituting the manual text input mechanism with a more convenient image upload mechanism that works seamlessly on mobile devices.
Solution Approach 2:
The patent introduces an intermediary system that automatically generates text queries from uploaded images. This intermediary process bridges the gap between image input and text-based search functionality, allowing users to benefit from both image convenience and text search capability without manual text entry.
2Ease of operation
If image-based query system is implemented, then user convenience is improved, but system complexity increases
Solution Approach 1:
The patent segments the image query system into distinct functional modules: image upload component, automatic text generation component, and search execution component. This segmentation allows each module to handle specific tasks independently, managing overall system complexity through modular design.
Solution Approach 2:
The patent employs intermediary components that translate between image data and text queries. These intermediaries handle the complex conversion processes, isolating the complexity from the user interface and allowing the core search functionality to remain relatively simple while supporting multiple query methods.
3Measurement precision
If automatic text generation from images is performed, then search accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-processing and analyzing images to generate text queries before the actual search execution. This preliminary text generation prepares the search parameters in advance, allowing the search process itself to proceed efficiently with pre-computed query terms.
Solution Approach 2:
The patent maintains continuity of useful action by overlapping the image analysis and text generation processes with the search execution process. Rather than completing all text generation before searching, the system continuously processes and refines queries during the search operation, maximizing resource utilization and reducing total processing time.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for processing queries made up of images. In one aspect, a method includes indexing images by image descriptors. The method further includes associating descriptive n-grams with the images. In another aspect, a method includes receiving a query, identifying text describing the query, and performing a search according to the text identified for the query.