Multi-Model Score Integration for Masked Face Authentication
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing information processing systems fail to effectively integrate and weigh the outputs of multiple inference models with different characteristics to prevent incorrect authentication results, particularly in biometric authentication scenarios such as face authentication with masks.
Innovation Solution
An information processing system that integrates scores from multiple inference models, including a first and second model with differing characteristics, calculates integrated scores with varying weights, and determines a specific state where only the second score exceeds a threshold, outputting appropriate information based on these conditions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a single inference model is used for authentication, then the system complexity is low, but the authentication accuracy and reliability deteriorate
Solution Approach 1:
The patent combines multiple inference models with different characteristics into a unified authentication system. The score integration unit merges results from multiple models, and the weight adjustment mechanism dynamically balances their contributions, achieving higher authentication reliability while managing system complexity through structured integration.
Solution Approach 2:
The system employs multiple inference models with different characteristics (e.g., different architectures, training data, or authentication criteria) to perform the same authentication function from different perspectives. This multi-functionality approach enhances reliability by cross-validating results while maintaining a universal authentication framework.
2Adaptability or versatility
If multiple inference models are integrated with equal weights, then the authentication coverage is improved, but the precision of score integration deteriorates
Solution Approach 1:
The patent implements dynamic weight adjustment in the score integration unit. Instead of using fixed equal weights, the system adaptively changes the weight of each inference model's score based on the specific authentication context, input characteristics, and model performance, thereby improving score integration precision while maintaining broad authentication coverage.
Solution Approach 2:
The system changes the parameter weights assigned to different inference models based on authentication requirements and model characteristics. The weight adjustment mechanism modifies these parameters dynamically to optimize the integration of scores from multiple models, achieving both broad coverage and high precision.
3Ease of operation
If the system outputs detailed guidance information for authentication failures, then the user interaction quality is improved, but the information processing complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the system analyzes authentication results from multiple inference models and provides targeted guidance information to users. The feedback unit generates specific recommendations based on which models failed and why, improving user interaction quality while managing processing complexity through selective analysis.
Solution Approach 2:
The system extracts and outputs only the necessary guidance information based on authentication failure patterns. Instead of processing and outputting all possible information, the system selectively extracts relevant guidance based on which inference models failed and the specific authentication context, reducing unnecessary processing complexity.
Data Source
AI summary
An information processing system includes: a score acquisition unit that acquires a first score based on an output of a first inference model and a second score based on an output of a second inference model; a first integrated score calculation unit that calculates a first integrated score; a second integrated score calculation unit that calculates a second integrated score; a score determination unit that determines whether or not each of the first integrated score and the second integrated score exceeds a predetermined threshold, and that determines whether or not it is in a specific state in which only the second integrated score exceeds the predetermined threshold; and an output unit that outputs an inference result in response to not being in the specific state, and that outputs information differing from the inference result in response to being in the specific state.