Image Processing Apparatus for Grouping Different Persons by Age

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image processing technologies fail to effectively find and group images of parents and children of the same age, despite advancements in person recognition and image search techniques.

Innovation Solution

An image processing apparatus and method that includes devices for recognizing and estimating the age of individuals in images, determining if they are the same person, and detecting image combinations where the individuals are not the same but have an age difference within a threshold range, allowing for the identification and grouping of images of relatives or affines captured at similar ages.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If person recognition techniques are used to identify the same person across images, then accurate person recognition is improved, but the ability to find images of different persons of the same age deteriorates

Engineering Contradiction:
Improveperson recognition accuracyVSAvoidability to find images of different persons of the same age
Core Design Contradiction:
Measurement precisionVSAdaptability or versatility

Solution Approach 1:

The patent segments the person recognition task into two distinct functions: (1) determining whether two recognized persons are the same person, and (2) determining whether the age difference between two persons is within a threshold range. This segmentation allows the system to simultaneously utilize person recognition accuracy while also identifying images of different persons of the same age, thereby resolving the contradiction between measurement precision and adaptability.

Inventive Principle:
Principle #1Segmentation

2Productivity

If image search techniques classify individual identification information for each person, then efficient image search is improved, but the ability to discover images of parents and children of the same age deteriorates

Engineering Contradiction:
Improveimage search efficiencyVSAvoidability to find images of parents and children of the same age
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent merges person identification information with age estimation information into a unified search framework. By combining these two types of information, the system maintains efficient image search capabilities while simultaneously enabling the discovery of images featuring parents and children of the same age, thus resolving the contradiction between productivity and adaptability.

Inventive Principle:
Principle #5Merging (Combining)

3Productivity

If image grouping is based on close imaging dates and times, then efficient image organization is improved, but the ability to find images of different persons captured in the same age deteriorates

Engineering Contradiction:
Improveimage organization efficiencyVSAvoidability to find images of different persons captured in the same age
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamic image grouping that can adapt between different grouping criteria. The system dynamically switches between time-based grouping (for efficient organization) and age-based grouping (for discovering images of different persons of the same age), thereby resolving the contradiction between productivity and adaptability in image organization.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentUS11475567B2Image processing apparatus, image processing method, image processing program, and recording medium storing program
Publication Date: 2022.10.18 FUJIFILM CORP
  • US11475567B2 patent drawing
  • US11475567B2 patent drawing
  • US11475567B2 patent drawing

AI summary

Provided are an image processing apparatus, an image processing method, an image processing program, and a recording medium that stores the program, capable of finding images of different persons captured in the same age.An image is specified as a first image, a user in the image is recognized as a first person, and an age thereof is estimated. An image is specified as a second image, a user in the image is recognized as a second person, and an age thereof is estimated. In a case where the user recognized as the first person and the user recognized as the second person are not the same person and a difference between the age of the user and the age of the user is within a range of a threshold value, the first image and the second image are detected as an image combination.