Raster Image Edge Tracing With Local Histogram Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for converting raster images to vector graphics face inaccuracies in edge detection, particularly in complex images, leading to excessive computer storage and interaction requirements due to erroneous edge detection and inability to handle edge directionality and illumination variations.
Innovation Solution
An edge detection system using a pixel window algorithm combined with a segmentation neural network to detect edges, employing local histogram rankings and object contours, with features like edge granularity determination, extrapolation, and one-click tracing to improve accuracy and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing image tracing systems generate vector graphics from raster images, then vectorization is achieved, but edge detection accuracy deteriorates in complex images leading to excessive anchor points and storage requirements
Solution Approach 1:
The patent segments the edge detection process into multiple independent modules: a pixel window algorithm for local edge analysis, a segmentation neural network for object contour detection, and an edge refinement system. This segmentation allows each module to specialize in specific aspects of edge detection, improving overall accuracy while enabling selective processing that reduces unnecessary data generation and storage requirements.
Solution Approach 2:
The patent dynamically adjusts detection parameters including pixel window size, edge strength thresholds, and neural network confidence levels based on image characteristics. By changing these parameters adaptively, the system optimizes edge detection accuracy for different image complexities while avoiding the generation of excessive anchor points that would increase storage requirements.
2Reliability
If existing systems trace all edges in raster images, then complete vectorization is achieved, but processing time and interaction requirements increase due to erroneous edge detection
Solution Approach 1:
The patent performs preliminary edge candidate identification using the pixel window algorithm before applying the more computationally intensive segmentation neural network. This preliminary action filters out obvious non-edge regions, allowing the neural network to focus only on ambiguous areas that require sophisticated analysis, thereby improving reliability while reducing overall processing time.
Solution Approach 2:
The patent applies full neural network processing only to regions where the pixel window algorithm detects ambiguous edge candidates, rather than processing the entire image uniformly. This partial application of computational resources maintains high detection reliability for critical regions while significantly reducing processing time for the overall image.
3Adaptability or versatility
If existing edge detection algorithms process entire images uniformly, then consistent processing is maintained, but flexibility in handling different image regions and illumination variations deteriorates
Solution Approach 1:
The patent implements local quality by allowing different processing strategies for different image regions. The pixel window algorithm adapts its parameters locally based on regional characteristics such as illumination conditions and edge density. The segmentation neural network similarly processes regions differently based on detected object boundaries, enabling the system to handle illumination variations and complex local features while maintaining manageable overall system complexity through modular design.
Data Source
AI summary
The present disclosure relates to systems, non-transitory computer-readable media, and methods for detecting and tracing edges in raster images using an advanced edge detection algorithm. For example, the disclosed systems generate, for a sliding pixel window in a raster image, a local histogram ranking of pixels according to pixel values located within the sliding pixel window. In some embodiments, the disclosed systems determine, for a region of the raster image enclosed by the sliding pixel window, an edge strength for an edge depicted in the raster image by comparing pixel ranks indicated by the local histogram ranking of the pixels. In certain embodiments, the disclosed systems also provide the edge for display based on determining the edge strength for the region of the raster image enclosed by the sliding pixel window.