Corner Pixel Seed Selection for Image Region Growing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional region growing techniques in computer imaging are inefficient due to the consideration of all surface points as potential seeds, leading to time-consuming processes, especially in large images with millions of pixels/voxels.
Innovation Solution
A method for selecting seeds from an image by determining a boundary between areas, filtering corner pixels, and placing them in a seed list for efficient region growing, where only specific seeds are chosen for growing regions, excluding those already part of a previously determined region.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all surface points are considered as potential seeds for region growing, then complete coverage of all regions is achieved, but the processing time becomes excessively long
Solution Approach 1:
The patent segments the boundary surface into distinct feature types (corner pixels, edge pixels, flat regions) and selectively processes only corner pixels as seeds. This segmentation allows the system to maintain complete region detection by focusing on the most informative boundary points while ignoring redundant areas, thereby resolving the contradiction between completeness and processing time.
Solution Approach 2:
The patent extracts and isolates corner pixels from the boundary surface using curvature analysis and discrete differential geometry operators. By extracting only these salient features as seed points, the system eliminates the need to process all surface points, achieving both complete region coverage and reduced processing time.
2Reliability
If traditional region growing considers every pixel/voxel as a potential seed, then no regions are missed, but the computational complexity increases significantly
Solution Approach 1:
The patent applies local quality by assigning different processing priorities to different boundary points based on their geometric properties. Corner pixels with high curvature are identified as important seeds, while flat or low-curvature regions are excluded. This local differentiation maintains detection completeness for all regions while significantly reducing computational complexity by focusing resources on informative locations.
Solution Approach 2:
The patent performs preliminary analysis of the boundary surface to identify and mark corner pixels before the region growing process begins. This preliminary action uses curvature computation and discrete differential operators to pre-select seed points, eliminating the need for complex real-time decisions during region growing and reducing overall computational complexity.
3Productivity
If corner pixels are selected as seeds, then processing efficiency is improved, but the method becomes more complex
Solution Approach 1:
The patent replaces complex geometric reasoning and manual seed selection with automated computational methods based on discrete differential geometry. By substituting mechanical/manual processes with mathematical operators that compute curvature and identify corner pixels algorithmically, the system achieves high processing efficiency while keeping the added complexity manageable through standardized mathematical operations.
Data Source
AI summary
An exemplary for selecting seeds from an image for region determination is provided. The method includes determining a boundary between two areas in the image; selecting pixels on the boundary that are characterized by a salient feature that identifies the pixels as seeds for determining a region; and determining a second region from one of the selected pixels if the one of the selected pixels is not part of a previously determined first region.