Preference-Weighted Still Image Generation From User History
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing photograph sales systems struggle to provide images that meet the preferences of users, particularly in childcare services, leading to difficulty in increasing sales.
Innovation Solution
An integrated ICT platform that includes a user management system, moving image distribution system, photograph generation system, and sales system, which analyzes user history and preferences to generate still images that align with individual user interests, using weight coefficients for 'Main Object Degree', 'Emotion Expression Degree', and 'Object to be Captured Together'.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Stability of the object's composition
If the technique suppresses variations among objects by finding a score in accordance with how well an object is captured, then the consistency of object capture quality is improved, but the ability to satisfy user preferences and increase sales deteriorates
Solution Approach 1:
The patent applies local quality by transitioning from uniform suppression of variations to targeted adjustment. Instead of treating all objects equally, the system identifies specific objects with high capture quality scores and selectively enhances them, while allowing other objects to maintain their original characteristics. This localized approach preserves overall consistency while introducing diversity to satisfy user preferences.
Solution Approach 2:
The patent implements dynamics by making the image processing parameters adjustable based on user preferences. The system dynamically modifies capture parameters and processing weights according to real-time user feedback and preference data, allowing the balance between consistency and diversity to shift adaptively rather than remaining fixed.
2Reliability
If the system provides photographs with uniform capture quality, then the reliability of image quality is improved, but the productivity of sales-generating content deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the image processing task into distinct stages: first evaluating capture quality scores, then selectively processing only the high-quality images that have potential to meet user preferences. This segmentation allows the system to maintain reliability for standard images while dedicating resources to enhancing sales-generating content separately.
Solution Approach 2:
The patent utilizes parameter changes by adjusting processing weights and capture parameters based on the evaluation scores. High-quality images receive enhanced processing weights to maximize their sales potential, while lower-quality images are processed with standard parameters. This dynamic parameter adjustment optimizes both reliability and productivity by allocating resources efficiently.
Data Source
AI summary
To make it possible to provide a photograph that meets the preference of a user. History information on an image purchased or viewed by a user and user information associating the user and a specific object with each other are obtained. Next, based on the obtained history information and the obtained user information, a weight for the preference of the user is derived. Then, based on the derived weight for the preference of the user, a still image that meets the preference of the user is generated.