Image Selection Using Relative Attribute Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods fail to effectively select and arrange images with desired attributes, often resulting in low probability of selecting intended images and high likelihood of similar images being grouped together, which hinders the portrayal of stories and inclusion of accent images with different looks.
Innovation Solution
An image processing apparatus and method that automatically selects and arranges images based on relative relationships between attribute values, such as focal length, shooting date/time, aperture, and resolution, allowing for varied image layouts and storytelling by setting conditions for frames in templates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If images having attribute information that matches attribute information set for a template are selected, then the selection process is automated, but the probability of selecting intended images becomes low
Solution Approach 1:
The patent changes the selection criterion from absolute attribute matching to relative relationship comparison. Instead of selecting images where attribute values exactly match template attributes (e.g., focal length = 100mm), the system selects images based on relative relationships (e.g., focal length of image A is greater than focal length of image B). This parameter transformation enables both automation and higher selection accuracy by comparing images against each other rather than against fixed template values.
2Ease of manufacture
If sub-images are extracted similarly and arranged on the same page as a main image, then the arrangement process is simplified, but similar images are arranged together making it difficult to portray a story
Solution Approach 1:
The patent applies local quality by assigning different selection criteria to different frame types. Main images are selected based on one set of attributes, while sub-images are selected based on relative relationships to the main image. This allows the system to maintain simple automated processing while achieving diverse and meaningful layouts, as each frame type has its own quality requirements that are locally optimized.
Solution Approach 2:
Instead of selecting sub-images with similar attributes to the main image (conventional approach), the patent inverts the logic by selecting sub-images based on relative differences and relationships. This inversion enables the system to automatically create varied layouts that tell stories, as sub-images are chosen to complement rather than replicate the main image.
3Ease of operation
If a focal length of 100 mm is set to select close-up images, then the selection criterion is simplified, but the probability of selecting intended close-up images becomes low
Solution Approach 1:
The patent transforms the selection parameter from an absolute value (focal length = 100mm) to a relative relationship (focal length of image A > focal length of image B). This change maintains operational simplicity by using the same focal length attribute but improves accuracy by comparing images relative to each other, thereby better identifying close-up images regardless of the specific focal length value.
Data Source
Figure 1
Figure 2~3
Figure 4
AI summary
An image processing apparatus (200) is provided including acquiring means (202) configured to acquire information regarding a set of frames that is associated with a condition indicating a relative relationship between a set of images to be arranged in the set of frames, extracting means (203) configured to extract the set of images from a plurality of images, wherein the set of images extracted satisfies the condition that is associated with the information regarding the set of frames acquired by the acquiring means (202), and arranging means (204) configured to arrange the set of images extracted by the extracting means (203) in the set of frames corresponding to the information acquired by the acquiring means (202).