Image Management Device Block Segmentation Matrix Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for ranking images based on importance degrees become resource-intensive as the number of images and objects increases, due to the large size of the probability transition matrix required for calculations, making it difficult for consumer devices to store and process effectively.
Innovation Solution
An image management device that groups images into blocks, calculates intra-block importance degrees, and then determines cluster and image importance degrees, reducing the resource load by using a smaller matrix size and normalizing importance degrees to minimize variations due to grouping methods.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a probability transition matrix is created based on the link structure between all image nodes and person nodes to calculate importance degrees, then the importance degree calculation becomes comprehensive and accurate, but the resource load grows exceedingly large with an increasing number of images and people
Solution Approach 1:
The patent segments the set of images into multiple blocks, where each block contains a subset of images. The link structure is constructed and importance degree calculation is performed separately for each block rather than for all images simultaneously. This segmentation reduces the size of the probability transition matrix from O(N^2) to O((N/B)^2) where B is the block size, significantly reducing computational resource requirements while maintaining calculation accuracy through the accumulation of importance degrees across blocks.
2Quantity of substance
If the number of images is increased to store more photographs, then the image storage capacity increases, but the difficulty to select important images increases due to the large number of stored images
Solution Approach 1:
The patent implements a feedback mechanism where the importance degree calculation is performed iteratively. The link structure is constructed from images and people, importance degrees are calculated for all nodes, and then these importance degrees are used to generate ranked lists of images and people. This feedback loop allows the system to automatically identify and prioritize important images based on their relationships with important people, making it easier for users to select relevant images from large collections.
3Measurement precision
If a link structure is created connecting images and people to calculate importance degrees, then the ability to identify important photographs improves, but the calculation complexity increases with the number of images and people
Solution Approach 1:
The patent divides the complete set of images into multiple blocks, where each block is processed independently to calculate importance degrees. This segmentation strategy reduces the calculation complexity from processing all images simultaneously to processing smaller subsets. The link structure is constructed within each block, and importance degrees are accumulated across blocks, achieving both accurate important photograph identification and reduced computational complexity.
Data Source
AI summary
An image management device clusters acquired images (S201) and generates blocks by grouping the images (S202). Next, the image management device calculates an intra-block importance degree of each cluster in each generated block (S204), calculates cluster importance degrees by accumulating the calculated intra-block importance degrees of each cluster (S205), and calculates an image importance degree based on the calculated cluster importance degrees (S206).