Hierarchical Database Segmentation for Fast Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current electronic devices face challenges in efficiently recognizing objects from image data due to high communication and processing resource usage, particularly in the context of augmented reality, where successful localization without feature classification is not possible without efficient matching of local feature descriptors.
Innovation Solution
The electronic device employs a processor to store and manage digital image data, using a cache or database to store reference image data based on usage conditions such as location or time, allowing for fast object recognition by loading relevant data for matching and updating the cache to optimize processing efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a large database is used to store reference image data for object recognition, then recognition accuracy is improved, but communication and processing resource usage increases
Solution Approach 1:
The patent divides the large database into multiple smaller databases organized in a hierarchical structure. The processor first searches smaller databases and only queries the larger database when necessary, segmenting the search process to reduce overall processing resource usage while maintaining recognition accuracy.
Solution Approach 2:
The patent performs preliminary actions by organizing and structuring the database beforehand into hierarchical levels. This pre-organization enables the processor to efficiently navigate and search through smaller subsets of data first, reducing the computational burden during actual object recognition operations.
2Measurement precision
If a large database is used to store reference image data, then object recognition accuracy is improved, but processing speed decreases
Solution Approach 1:
The database is segmented into hierarchical levels with smaller databases at lower levels and a larger database at higher levels. The processor implements a segmented search strategy, querying smaller databases first and progressively moving to larger databases only when necessary, thereby maintaining high recognition accuracy while improving processing speed.
Solution Approach 2:
The database structure is preliminarily organized into hierarchical levels before operation. This pre-established structure enables the processor to quickly determine which database level to search based on the recognition task, avoiding unnecessary searches through the entire large database and thus improving recognition speed.
3Reliability
If extensive database searches are performed for object recognition, then recognition completeness is improved, but resource consumption increases
Solution Approach 1:
The search process is segmented into multiple stages corresponding to different database levels. The processor performs exhaustive searches within smaller databases first, ensuring recognition completeness for common objects, and only proceeds to search larger databases when objects are not found, thus reducing overall resource consumption while maintaining completeness.
Solution Approach 2:
The hierarchical database structure is preliminarily configured to optimize search efficiency. This pre-arranged structure allows the processor to systematically navigate through databases in an order that ensures completeness while minimizing unnecessary resource consumption by avoiding searches in larger databases when smaller ones suffice.
Data Source
AI summary
An electronic device for fast object recognition is provided. The electronic device includes a first storage unit configured to store digital image data, and a processor configured to recognize an object in first image data, to receive a second object related to the first object in the first image data from a second storage unit, to store the first and second objects in the first storage unit, and to use one or more of the first and second objects stored in the first storage unit to recognize an object in second image data.