Line Segment Detector for Image Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional line segment detectors only extract the contour of characters or lines, leading to poor compression efficiency and image quality, as they fail to detect the content within the character area, and do not correct erroneous detections such as isolated points, resulting in noise components in image processing and compression.
Innovation Solution
A line segment detector and method that re-determines the detection of line segments by counting pixels within a window, using threshold values based on pixel averages, and performs pre-processing to correct erroneous detections, expanding the detection to include both the contour and content of characters, and removing noise components like isolated points.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If only contour detection is performed using conventional line segment detectors, then the detection process is simple and fast, but the compression efficiency and image quality deteriorate because the content within character areas is not extracted
Solution Approach 1:
The detection process is divided into two stages: first detecting line segments (contours), then performing a second detection to identify character areas. This segmentation allows the system to extract both contour and content information, improving compression efficiency without overwhelming complexity in a single detection pass
Solution Approach 2:
The first line segment detection is performed as a preliminary step to identify contours, which then serves as input for the second detection stage. This preliminary action enables the system to focus subsequent processing on character area identification, achieving both contour and content extraction
2Reliability
If conventional line segment detection is used, then the detection method is simple, but erroneous detections such as isolated points occur, resulting in noise components that deteriorate image quality
Solution Approach 1:
The second line segment detection acts as a feedback mechanism that re-evaluates pixels detected in the first stage. By counting detected pixels within windows and comparing against thresholds, the system corrects erroneous detections like isolated points, improving reliability with a manageable correction mechanism
Solution Approach 2:
The second detection stage performs preliminary anti-action by identifying and removing erroneous detections before final image processing. By setting thresholds and counting pixels in advance, the system prevents noise components from affecting the final output
3Area of stationary object
If the detection window size is increased to capture more content, then the content extraction improves, but the detection precision for line segments decreases
Solution Approach 1:
The detection process uses different window sizes for different purposes: smaller windows for line segment detection to maintain precision, and larger windows for character area detection to capture content. This segmentation of detection parameters allows both precision and coverage to be optimized for their respective functions
Solution Approach 2:
The second detection stage uses a larger window size that may be excessive for precise line segment detection, but this is acceptable because the primary goal at this stage is to identify character areas. The system accepts some loss in line segment precision to achieve comprehensive character area coverage
Data Source
AI summary
A line segment detector comprises a vertical line detecting part, a horizontal line detecting part, a left-to-right downward slanting line detecting part, a left-to-right upward slanting line detecting part, and a line-segment detecting/synthesizing section that calculates OR operation of output values from the vertical line detecting part, the horizontal line detecting part, the left-to-right downward slanting line detecting part, and the left-to-right upward slanting line detecting part. The line segment detector further comprises a re-determining unit comprising a second line-segment detecting part that sets a window to the detection result near the target pixel, counts the number of pixels whose detection result is determined as the line segment in the set window, determines, when the counted value is not smaller than a predetermined threshold value, that the pixel is a line segment.