Image Processing Apparatus Dynamic Threshold Character Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods face difficulties in detecting characters in large rectangular areas efficiently due to the need for complex calculations, leading to slow processing times and potential misdetection.
Innovation Solution
An image processing apparatus comprising an edge image generating unit, connection pixel extracting unit, binary image generating unit, background image generating unit, and image compressing unit, which detects edges, extracts connection pixel sets, classifies them by color, and uses color difference thresholds to identify characters, adjusting threshold values based on rectangular area size for accurate and fast character detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If character detection is performed using labeling rate or circumference length rate for large rectangular areas, then detection accuracy is improved, but processing time increases significantly
Solution Approach 1:
The patent segments the character detection process into two distinct paths: one for small rectangular areas using simple area comparison, and another for large rectangular areas using labeling rate and circumference length rate. This segmentation allows the system to apply computationally intensive methods only when necessary, reducing overall processing time while maintaining accuracy for complex cases.
Solution Approach 2:
The patent changes the detection parameters based on the size of the rectangular area. For small areas, it uses area threshold comparison; for large areas, it switches to labeling rate and circumference length rate calculations. This dynamic parameter adjustment optimizes the balance between detection accuracy and processing efficiency for different input scenarios.
2Productivity
If simple area-based detection is used for small rectangular areas, then processing speed is improved, but detection accuracy may deteriorate
Solution Approach 1:
The patent dynamically changes detection parameters based on rectangular area size. Small areas use simple area threshold comparison for fast processing, while large areas trigger more accurate but computationally intensive labeling rate and circumference length rate calculations, ensuring both speed and accuracy are optimized for their respective scenarios.
Solution Approach 2:
The detection method is made dynamic by adjusting the complexity of the detection algorithm based on the input characteristics (rectangular area size). The system transitions between different detection strategies rather than using a fixed approach, allowing it to adapt processing intensity to the actual needs of each detection task.
3Device complexity
If fixed color difference threshold is used for character identification, then processing simplicity is maintained, but detection accuracy deteriorates for varying size rectangular areas
Solution Approach 1:
The patent changes the color difference threshold parameter based on the size of the rectangular area. Small rectangular areas use one threshold value, while large rectangular areas use a different threshold value. This dynamic parameter adjustment maintains processing simplicity while significantly improving character identification accuracy across different input scenarios.
Data Source
Figure 1
Figure 2A~2C
Figure 3
AI summary
Regarding a color difference between a color of an original image on a pixel other than a connection pixel set in a rectangular area that is an inside area of a rectangle circumscribed to the connection pixel set and a color of the original image on a pixel in the connection pixel set, the binary image generating unit (4) (a) identifies whether the connection pixel set is a character or non character by comparing a value of an index that indicates unevenness of the color differences with a color difference threshold value, (b) identifies whether a size of the rectangular area is small or large on the basis of a threshold value, and (c) sets the color difference threshold value as a value if the size of the rectangular area is small and as another different value if the size of the rectangular area is large.