Recommendation Control Device Face Authentication Noise Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing recommendation systems face a decrease in accuracy due to noise and time zone variations in user behavior histories, leading to unsuitable recommendation information.
Innovation Solution
A recommendation control device and system that captures images, extracts face features, performs face authentication, and extracts relevant behavior histories to identify personalized recommendation information based on predetermined conditions, ensuring accurate and individualized recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all individual behavior histories are used to generate recommendation information, then the quantity of recommendation data increases, but recommendation accuracy decreases due to noise and time zone variations
Solution Approach 1:
The patent extracts only the necessary and relevant behavior history data that meets predetermined conditions, rather than using all available behavior histories. This selective extraction removes noise and irrelevant information that would degrade recommendation accuracy while maintaining sufficient data quantity for personalized recommendations.
Solution Approach 2:
The patent applies different processing qualities to different portions of behavior history data by setting predetermined conditions for extraction. Relevant behavior histories meeting specific criteria are extracted with high quality for recommendation generation, while irrelevant data is excluded, creating a differentiated quality approach that improves overall recommendation precision.
2Adaptability or versatility
If behavior histories from different time zones and situations are included, then the versatility of recommendation data increases, but the reliability of recommendation information decreases
Solution Approach 1:
The patent segments behavior history data into distinct categories based on predetermined conditions such as time zones and situations. By dividing the overall behavior history into manageable segments and selectively extracting those that meet specific criteria, the system maintains versatility across different user contexts while ensuring reliability through condition-based filtering of relevant segments.
Data Source
AI summary
A recommendation control device for providing recommendation information that suits an individual preference is provided. The recommendation control device (100) includes: an acquisition unit (130) that acquires a captured image being captured by a predetermined photographing device; an authentication control unit (140) that extracts a face area or face feature information from the captured image, and causes an authentication device (200) to perform face authentication; a behavior history extraction unit (150) that extracts a behavior history that satisfies a predetermined extraction condition from a behavior history of a user successful in the face authentication; an identification unit (160) that identifies recommendation information, based on a behavior history extracted by the behavior history extraction unit (150); and an output unit (170) that transmits recommendation information identified by the identification unit (160) to a predetermined display terminal.