Face Group Association for Name Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for associating names and face images, such as those in Patent Literature 2, face difficulties when the person's face cannot be detected or when individuals are not major characters in programs, leading to inefficient registration processes.
Innovation Solution
An information processing device that selects content based on user-specified names, groups face images per person, and determines the appropriate face group to associate with the name using similarity patterns and occurrence frequencies, enabling efficient registration even if the name is not initially registered.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual registration of names and face images is conducted, then registration accuracy is improved, but registration efficiency deteriorates
Solution Approach 1:
The system enables automatic registration by having the registration unit autonomously extract names from EPG data, detect face images from program content, and associate them without requiring manual intervention. The system serves itself by automatically gathering necessary data from multiple sources and performing the association task independently, thus resolving the contradiction between accuracy and efficiency.
Solution Approach 2:
The manual mechanical process of registration is replaced with an automated information processing system that uses data extraction, face detection algorithms, and automatic association mechanisms. This substitution of manual operations with automated computational processes maintains registration accuracy while dramatically improving efficiency.
2Productivity
If face detection is used to associate names and face images, then registration speed is improved, but reliability deteriorates when face cannot be detected
Solution Approach 1:
The registration unit performs multiple functions: it attempts face detection for quick association, but also extracts names from EPG data and uses this alternative information source when face detection fails. This multi-functional approach ensures that registration can proceed reliably through different methods depending on what data is available, maintaining both speed and reliability.
Solution Approach 2:
The system preliminarily extracts names from EPG data and prepares this information before face detection is attempted. This preliminary preparation of name data ensures that when face detection fails, the association can still proceed using the pre-extracted name information from program metadata, thus maintaining reliability while preserving registration speed.
3Extent of automation
If EPG data is used to extract names, then automation level is improved, but adaptability deteriorates for non-major characters and news programs
Solution Approach 1:
The registration unit merges multiple data sources: EPG information for automated name extraction, face detection results from program content, and additional metadata. By combining these diverse information sources, the system maintains high automation while improving adaptability to capture names of non-major characters and individuals in news programs that may not appear in standard EPG listings.
Data Source
AI summary
An information processing device that includes circuitry that selects content including at least one name specified by a user; selects a plurality of contents in which the name possibly occurs; acquires a plurality of face groups by grouping, per person, face images occurring in the plurality of contents selected by the circuitry; calculates a number of similar face groups appearing in other content for each of the plurality of face groups; acquires, as a candidate face group, a face group having a greatest number of similar face groups appearing in other content; and determines a face group to associate with the name from face groups.