Image Processing Apparatus Tangent Projection for Bounding Rectangle
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for computing minimum bounding rectangles in multiple directions suffer from increased computational complexity, which limits estimation accuracy in real-time systems with limited resources.
Innovation Solution
An image processing apparatus and method that extracts contours, calculates tangents, projects these tangents orthogonally, and computes rectangles using intersections with maximum and minimum values of coordinates, thereby reducing computational complexity by limiting tangent directions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of directions for computing minimum bounding rectangles is increased to improve estimation accuracy, then the direction estimation accuracy of character rows is improved, but the computational complexity increases
Solution Approach 1:
The patent segments the computational process into two distinct phases: (1) computing minimum bounding rectangles only in the initial horizontal and vertical directions to obtain candidate regions, and (2) performing direction estimation only within those candidate regions. This segmentation avoids computing minimum bounding rectangles in all possible directions, thereby reducing computational complexity while maintaining estimation accuracy.
Solution Approach 2:
The patent extracts and processes only the most relevant candidate regions identified from initial horizontal and vertical minimum bounding rectangle computations. By taking out only these specific candidate regions for further direction estimation processing, the system avoids the computational burden of processing all regions in all directions, thus reducing overall computational complexity while preserving accuracy for relevant areas.
2Measurement precision
If the number of directions for computing minimum bounding rectangles is increased, then the direction estimation accuracy is improved, but the processing time increases
Solution Approach 1:
The patent divides the processing into stages: first computing minimum bounding rectangles in only two directions (horizontal and vertical) to quickly identify candidate regions, then performing detailed direction estimation only on those candidates. This staged approach significantly reduces processing time compared to computing minimum bounding rectangles in all directions, while still achieving accurate direction estimation for character rows.
Solution Approach 2:
The patent performs preliminary computation of minimum bounding rectangles in horizontal and vertical directions before the final direction estimation step. This preliminary action identifies and narrows down candidate regions in advance, so that the computationally intensive direction estimation is performed only on a small subset of regions, thereby reducing total processing time while maintaining accuracy.
3Measurement precision
If the number of directions for computing minimum bounding rectangles is increased, then the direction estimation accuracy is improved, but the computational load increases
Solution Approach 1:
The patent segments the computational workload by computing minimum bounding rectangles in only two directions (horizontal and vertical) rather than all possible directions. This segmentation reduces the computational load and energy consumption while still providing sufficient candidate regions for accurate direction estimation in subsequent processing steps.
Solution Approach 2:
The patent applies partial action by computing minimum bounding rectangles only in the necessary horizontal and vertical directions rather than in all possible directions. This partial computation approach reduces computational load and energy usage while obtaining sufficient information for accurate direction estimation through the two-stage processing method.
Data Source
AI summary
An image processing apparatus include: a storage unit storing an image of a processing target; a tangent calculating unit extracting contours as a bent lines represented by sets of contour points from an image read from the storage unit and computing tangents to the extracted contour; a projecting unit projecting computed tangents to axes in directions orthogonal to the corresponding tangents, and computing coordinates of intersections where the tangents intersect the axes; and a rectangle calculating unit selecting intersections with maximum values and minimum values of coordinates among intersections computed by the projecting unit for each direction of the axis, and computing a rectangle formed by a pair of parallel tangents passing through two intersections with maximum values and minimum values selected for a first axis and another pair of tangents passing through two intersections with maximum values and minimum values selected for a second axis orthogonal to the first axis.