Face-Authenticated Group Recommendations by Action and Attribute
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing recommendation technologies fail to provide accurate recommendations due to the variability in group actions despite similar group attributes.
Innovation Solution
A system that performs face authentication on detected face regions, determines group authentication based on successful face authentication, specifies group actions, estimates group attributes, and selects appropriate recommendations based on these attributes for output to display terminals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If group attributes are used for recommendation, then recommendation coverage is improved, but recommendation accuracy deteriorates because effective recommendation information differs depending on actions of the groups
Solution Approach 1:
The patent segments the recommendation approach by introducing a new dimension of classification: not only group attributes but also group actions. The determination unit divides groups into first groups (with at least one member who performed a predetermined action) and second groups (without such members), then applies different recommendation strategies to each segment, thereby improving accuracy while maintaining coverage.
Solution Approach 2:
The system dynamically adjusts recommendation content based on real-time detection of group actions. The determination unit continuously monitors whether members of a group have performed predetermined actions, and the output unit dynamically selects different recommendation information accordingly. This dynamic adaptation allows the system to maintain high accuracy by responding to actual group behavior rather than relying solely on static attributes.
2Measurement precision
If face authentication is performed for all detected face regions, then group identification accuracy is improved, but processing time increases
Solution Approach 1:
The patent applies partial action by performing face authentication selectively rather than universally. The personal authentication control unit determines whether to perform authentication based on the captured image content and group characteristics. This selective approach maintains sufficient identification accuracy for most cases while significantly reducing processing time by avoiding unnecessary authentication operations.
Solution Approach 2:
The system incorporates feedback mechanisms where the determination unit uses authentication results to adjust subsequent processing steps. When authentication succeeds for certain members, the system uses this feedback to determine group composition and action status, thereby optimizing the authentication process and reducing redundant processing time.
Data Source
AI summary
A recommendation device includes: an acquisition unit configured to acquire a captured image; a personal authentication control unit configured to control to perform face authentication for a plurality of face regions detected from the captured image; a determination unit configured to determine that group authentication has succeeded in a case where face authentication using at least one of the plurality of users registered as a group and any one of the plurality of detected face regions has succeeded; a specifying unit configured to specify an action in the group based on the point in a case where the group authentication has succeeded; an estimation unit configured to estimate a group attribute based on a personal attribute of each user belonging to the group; a selection unit configured to select recommendation information based on the group attribute and the action; and an output unit configured to output the selected recommendation information.