Face Recognition Correction via Dual Management Units
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for automatically grouping faces in large image datasets struggle with accuracy in recognizing and distinguishing between individuals, leading to misrecognition issues.
Innovation Solution
An apparatus and method that includes a first management unit for classifying feature information, a second management unit for managing objects, an association unit for linking objects with feature information, and an input unit for correcting classifications, allowing for improved recognition accuracy by updating the classification based on user input.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic face grouping is performed using existing methods, then the processing speed is improved, but the recognition accuracy deteriorates leading to misrecognition
Solution Approach 1:
The system implements feedback by detecting misrecognition cases during automatic grouping and using warning marks to alert users. When users correct misrecognized faces, this correction information is fed back to improve the grouping algorithm, creating a continuous improvement loop that maintains both speed and accuracy
Solution Approach 2:
The patent segments the face grouping process into multiple stages: initial automatic grouping for speed, followed by selective manual verification for accuracy-critical cases. This segmentation allows the system to process most faces automatically while dedicating human attention only to ambiguous or potentially misrecognized cases
2Measurement precision
If manual verification of each face is performed to improve recognition accuracy, then the recognition accuracy is improved, but the time consumption increases
Solution Approach 1:
Instead of requiring complete manual verification of all faces, the system applies partial manual action only to cases where misrecognition is detected or suspected. The warning mark mechanism identifies specific faces needing attention, so users perform verification only when necessary, reducing overall time consumption while maintaining accuracy
3Measurement precision
If warning marks are displayed for potential misrecognition to improve accuracy, then the recognition accuracy is improved, but the user interface complexity increases
Solution Approach 1:
The system extracts only the critical information needed for accuracy improvement - specifically, displaying warning marks only on faces that are likely misrecognized. This selective extraction avoids overwhelming the user with unnecessary information while still providing the accuracy benefits of identifying problematic cases
Data Source
AI summary
An apparatus includes a first management unit configured to classify and manage feature information of a plurality of objects extracted from image data in units of similar feature information, a second management unit configured to classify and manage the plurality of objects extracted from the image data object by object, an association unit configured to associate the objects in the first management unit with feature information in the second management unit, and an input unit configured to input a correction instruction about classifications of the objects by the second management unit. The second management unit is configured to, if the correction instruction is input, correct and manage a classification to which a target object of the correction instruction belongs and the first management unit is configured to change and manage a classification to which feature information associated with the target object of the correction instruction belongs.