Intimacy Calculation in Image Processing Apparatus
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing systems fail to accurately calculate the intimacy between persons appearing in images, as they do not consider the actual relationships between individuals who often act as photographers but may not appear frequently in images.
Innovation Solution
An image processing apparatus that acquires images from multiple users, analyzes the photographer information, and calculates intimacy based on the presence and frequency of individuals in images, considering factors like the number of images, facial expressions, and locations, to create a composite image reflecting actual relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If intimacy is calculated based only on image contents (persons appearing in images), then calculation is simple, but accuracy deteriorates because photographers who are intimate with persons may not appear in images
Solution Approach 1:
The patent merges two intimacy calculation methods: one based on image contents (persons appearing together) and another based on photographer information (who captured the images). By combining these approaches, the system achieves more accurate intimacy calculation that accounts for both visible interactions and the photographer's relationship with subjects, resolving the contradiction between simplicity and accuracy.
2Measurement precision
If multiple factors are considered for intimacy calculation (photographer info, image contents, frequency), then accuracy improves, but processing complexity increases
Solution Approach 1:
The patent segments the intimacy calculation process into distinct modules: one analyzing image contents (detecting persons, objects, scenes), another analyzing photographer information (metadata, capture frequency), and a third integrating these factors with weighted scoring. This segmentation allows complex multi-factor analysis to be performed in manageable, independent steps, reducing overall processing complexity while maintaining accuracy.
Solution Approach 2:
The system dynamically adjusts the weight of different factors based on the specific context. For example, if photographer information is reliable and available, it receives higher weight; if image contents provide stronger signals, those are emphasized. This dynamic weighting allows the system to adapt to varying data quality and availability, managing complexity through flexibility rather than rigid fixed-weight approaches.
3Measurement precision
If all images are analyzed in detail for intimacy calculation, then accuracy improves, but processing time increases
Solution Approach 1:
The patent applies partial analysis by focusing computational resources on the most informative images rather than analyzing every image in equal detail. It identifies key images based on criteria such as capture frequency, person appearance prominence, and photographer consistency, then performs detailed analysis on these selective images while using simplified analysis for others. This partial action approach maintains accuracy for critical cases while significantly reducing overall processing time.
Data Source
AI summary
There are provided an image processing apparatus, an image processing method, a program, and a recording medium capable of calculating the actual intimacy between persons even in a case where there is a deviation in the actual intimacy between persons by calculating the intimacy between persons in consideration of only the contents of images. In the image processing apparatus, the image processing method, the program, and the recording medium of the invention, a person specifying unit determines persons appearing in each image, and specifies one or more first persons among the determined persons. An intimacy calculation unit calculates the intimacy between the first person and each photographer based on the presence of a first person in each image and photographer information of each image.