Pattern Recognition Device Candidate Re-ranking Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current pattern recognition devices face challenges in achieving accurate recognition results due to the ranking of recognition candidates being fixed based on initial recognition scores, which can lead to incorrect identification of the correct answer data when multiple candidates have similar scores.
Innovation Solution
A pattern recognition device and learned model that calculate an evaluation value for each pair of recognition candidates to assess the possibility of the second candidate being the correct answer, allowing for re-ranking and improving recognition accuracy by determining the final recognition result based on these evaluation values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If recognition candidates are ranked based on recognition scores from multiple recognition schemes, then recognition efficiency is improved, but recognition accuracy deteriorates because the ranking does not change between candidates with different scores
Solution Approach 1:
The patent transforms the single-dimensional recognition score into a multi-dimensional evaluation system by introducing correction values derived from relationships between recognition candidates. This changes the parameter structure from simple score-based ranking to a composite evaluation incorporating candidate relationships, thereby improving accuracy while maintaining efficiency.
Solution Approach 2:
The patent introduces an intermediary evaluation mechanism that assesses the relationship between recognition candidates (particularly between the highest-scoring candidate and other candidates). This intermediary evaluation acts as a mediator to adjust the final ranking, ensuring that the correct answer is not overlooked due to rigid score-based ordering.
2Device complexity
If the ranking of recognition candidates is maintained based on initial recognition scores, then the processing complexity is reduced, but the ability to identify correct answer data deteriorates when candidates have similar scores
Solution Approach 1:
The patent performs preliminary evaluation of the relationship between recognition candidates before final ranking determination. By calculating correction values based on candidate relationships in advance, the system prepares adjustment information that can be applied to improve identification reliability without significantly increasing overall processing complexity.
Solution Approach 2:
The patent introduces dynamic adjustment to the static score-based ranking system. The final ranking is not fixed but dynamically adjusted based on the evaluated relationships between candidates. This allows the system to adapt the ranking when necessary to improve correct answer identification while maintaining simplicity in normal cases.
Data Source
AI summary
A pattern recognition device (10) includes an acquisition unit (13) configured to acquire a plurality of recognition candidates obtained by pattern recognition for an input image and having a recognition score, a calculation unit (17) configured to calculate an evaluation value obtained by quantifying a possibility of a recognition candidate (K) being correct answer data corresponding to the input data on the basis of a feature quantity regarding a recognition candidate (A1) and the recognition candidate (K), for each pair of the recognition candidate (A1) having the number one recognition score and another recognition candidate (K) among the plurality of recognition candidates, and a determination unit (18) configured to determine a final recognition result from among the plurality of recognition candidates on the basis of an evaluation value for each recognition candidate (K).