Dual OCR Recognition Probability Thresholding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
When recognition results from two OCRs match, but the recognition probability is low, the output may not be the correct answer, leading to increased human workload for verification.
Innovation Solution
An information processing apparatus that acquires recognition results and probabilities from two OCRs, checks for matching results, and applies different processing based on threshold comparisons of the recognition probabilities to determine the reliability of the output, reducing human workload by only verifying results with high probabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If recognition results from two OCRs are matched and output without probability checking, then productivity is improved, but reliability deteriorates because low-probability matches may be incorrect
Solution Approach 1:
The patent introduces probability thresholds as a parameter to control the output of recognition results. By comparing the recognition probability against a threshold value, the system dynamically determines whether to output the matched result automatically or require manual verification, thus maintaining reliability while preserving productivity for high-confidence cases.
2Reliability
If all matched recognition results are verified manually, then reliability is improved, but productivity deteriorates due to increased human workload
Solution Approach 1:
The patent applies different processing qualities to different recognition results based on their probability values. High-probability results receive automatic output treatment, while low-probability results receive manual verification treatment. This local differentiation ensures reliability is improved only where necessary, preserving overall productivity.
3Reliability
If probability threshold checking is implemented, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent implements probability threshold checking only for matched recognition results, not for all processing paths. This partial application of the checking mechanism provides reliability improvement where needed (for matched results) while avoiding unnecessary complexity in other processing branches, thus managing device complexity effectively.
Data Source
AI summary
An information processing apparatus includes a processor configured to acquire a first recognition result and a first recognition probability on target data from a first recognizer, acquire a second recognition result and a second recognition probability on the target data from a second recognizer, execute checking of the first recognition result and the second recognition result, and execute first control in a case where the first recognition result and the second recognition result match each other as a result of the checking. The first control is control for executing either of first processing or second processing on the matched recognition result and outputting a processing result based on at least one of the first recognition probability or the second recognition probability. A human workload for the first processing is smaller than a human workload for the second processing.