Combined Image Selection Using Similarity-Aware Quality Assessment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods struggle to generate high-quality combined images from multiple images due to loss of color balance and viewer appeal when images are arranged together, especially when similar images are included, leading to a low-quality combined image.
Innovation Solution
An information-processing device that selects and arranges images in predefined frames using a combination module, with an assessment module determining a total assessment value based on selection, combination, and similarity values, employing machine learning for image quality evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If images are automatically selected and arranged based on certain criteria, then generation efficiency is improved, but image quality and viewer appeal deteriorate due to loss of color balance and similarity issues
Solution Approach 1:
The patent implements a feedback mechanism where the generated combined image is assessed using an assessment module that calculates a total assessment value based on color balance, image quality, and similarity metrics. This feedback is used to iteratively improve the selection and arrangement of images, allowing the system to learn from previous generations and progressively enhance image quality while maintaining high generation efficiency through automated optimization.
Solution Approach 2:
The system dynamically adjusts the selection and arrangement of images based on real-time assessment of color balance, similarity, and overall quality. Rather than using fixed criteria, the system adapts its selection strategy by evaluating the impact of each image addition on the combined image's overall quality, enabling it to optimize both efficiency and quality simultaneously through dynamic decision-making.
2Quantity of substance
If a large number of images are arranged in the combined image, then content richness is improved, but color balance and overall quality deteriorate
Solution Approach 1:
The patent changes the parameters of image selection by introducing multiple assessment dimensions including color balance metrics, similarity measures, and quality scores. The system evaluates potential image combinations based on these parameters and selects arrangements that optimize the balance between quantity of images and overall quality, using mathematical models to determine the optimal number and combination of images for the combined image.
3Stability of the object's composition
If similar images are included in the combined image, then content consistency is improved, but viewer appeal and quality deteriorate due to peculiar feelings
Solution Approach 1:
The patent applies local quality control by assessing similarity between individual images in the combined image and adjusting the selection to ensure diversity where needed. The system calculates similarity metrics between image pairs and uses this information to prevent excessive similarity in the final composition, while still maintaining thematic consistency through controlled selection criteria that balance local diversity with global coherence.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Provided is an information-processing device (100) including: a single image selection module (10) configured to select M (M≤N) images from N (N>1) images; a combination module (20) configured to generate a combined image by arranging the selected M images in M frames defined in advance, respectively; and an assessment module (30) configured to determine a total assessment value in association with the generated combined image, the total assessment value including at least a linear sum of a selection assessment value being a linear sum of single image assessment values of the selected M images and a combination assessment value being a single image assessment value of the combined image.