Dynamic Vector Distance Threshold for Character Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing character recognition systems face challenges in simultaneously updating and comparing data libraries while performing character recognition, leading to inefficiencies and potential misjudgments due to static threshold settings.
Innovation Solution
A character recognition method and device that determines matching characters by calculating vector distances between image vectors and registered character vectors, storing matching vector distances, and dynamically updating data using a moving average method to adjust the vector distance threshold based on historical data, ensuring smoother trend maintenance and reducing misjudgments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a static vector distance threshold is used for character recognition, then the recognition process is simple and fast, but misjudgments occur due to inability to adapt to data changes
Solution Approach 1:
The patent transforms the static threshold into a dynamic one by introducing a moving average mechanism. The threshold is continuously updated based on historical recognition data, allowing it to adapt to changes in the data distribution over time. This dynamic adjustment resolves the contradiction by maintaining high recognition accuracy through adaptability while keeping the implementation relatively simple through the use of averaging calculations.
Solution Approach 2:
The patent implements a feedback loop where recognition results are fed back into the system to update the threshold. The moving average of historical vector distances is continuously calculated and used to adjust the threshold, creating a self-regulating mechanism that improves reliability without requiring complex external control systems.
2Reliability
If the data library is updated frequently to improve recognition accuracy, then recognition reliability improves, but system performance deteriorates due to simultaneous read-write operations
Solution Approach 1:
The patent performs preliminary actions by pre-calculating and storing the moving average threshold separately from the main recognition workflow. The threshold is updated in the background based on historical data, so that during actual recognition operations, the pre-computed threshold can be quickly applied without causing write conflicts or performance degradation.
Solution Approach 2:
The patent introduces the moving average threshold as an intermediary between the raw recognition data and the decision-making process. This intermediary absorbs the complexity of frequent updates by aggregating historical data into a single threshold value, allowing the system to maintain high reliability while avoiding the performance penalty of frequent direct updates to the recognition logic.
3Reliability
If instantaneous changes are filtered out using moving average method, then misjudgments are reduced, but the system responds slower to legitimate changes
Solution Approach 1:
The patent adjusts the parameter of the moving average window size to balance filtering effectiveness and response speed. By optimizing the window size, the system can filter out short-term fluctuations and noise while still responding adequately to genuine changes in character recognition patterns. This parameter tuning resolves the contradiction between smoothing out misjudgments and maintaining responsiveness.
Data Source
AI summary
A character recognition method includes the following operations: determining that the image of character to be identified corresponds to a matching character of several registered characters according to several vector distances to be identified between a vector of an image of character to be identified and several vectors of several registered character images of several registered characters, and storing a matching vector distance between the vector of the image of character to be identified and a vector of the matching character by a processor; and storing a data of the matching character according to the image of character to be identified when the matching vector distance is less than a vector distance threshold by the processor.