Image Segmentation Using Saliency Model for Sample Point Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image segmentation methods require manual selection of foreground and background sample points by users, leading to low efficiency when processing large numbers of images.
Innovation Solution
An image segmentation method that establishes a saliency model to automatically determine foreground and background sample points, combining these with a graph cut algorithm for image segmentation using edge information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual selection of foreground and background sample points is used, then segmentation accuracy can be maintained, but processing efficiency deteriorates when handling large numbers of images
Solution Approach 1:
The system automatically selects foreground and background sample points using a saliency model without requiring manual user input. The saliency model independently identifies salient regions and automatically determines sample points, enabling the system to serve itself rather than relying on manual operation, thus improving processing efficiency while maintaining segmentation accuracy
Solution Approach 2:
The patent introduces a saliency model that changes the parameter selection approach from manual coordinate specification to automated saliency-based sampling. By computing saliency maps and selecting sample points based on saliency thresholds, the system transforms the sampling process into an automated parameter-driven operation, resolving the contradiction between accuracy and efficiency
2Productivity
If automatic sample point selection using saliency model is implemented, then processing efficiency improves, but segmentation accuracy may deteriorate without manual verification
Solution Approach 1:
The system incorporates a feedback mechanism where the saliency model continuously refines sample point selection based on segmentation results. The graph cut algorithm uses the selected sample points to generate segmentation, which then feeds back to validate and potentially adjust sample point selection, ensuring accuracy is maintained while preserving automated efficiency
Solution Approach 2:
The saliency model performs preliminary analysis to pre-select high-quality sample points before the actual segmentation process. By conducting this preliminary action of identifying salient regions and selecting representative sample points in advance, the system ensures that the subsequent segmentation algorithm receives optimal input, maintaining accuracy while enabling automated high-speed processing
Data Source
AI summary
The present disclosure provides an image segmentation method and device. The image segmentation method comprises: establishing a saliency model of an image; obtaining foreground sample points and background sample points of the image according to the saliency model; establishing a foreground and background classification model according to the saliency model, the foreground sample points and the background sample points; and segmenting the image according to a predefined graph cut algorithm which segments the image by using the foreground and background classification model and edge information related to pixels in the image. A problem that a user needs to manually and roughly select the foreground sample points and the background sample points and thus the segmentation efficiency is low for segmenting a large number of images is solved.