Image Segmentation for Contacting Objects Separation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods struggle to accurately classify and quantify objects in images, especially when objects are in contact with each other or partially overlap.
Innovation Solution
An image processing apparatus with modules for inputting images, segmenting foreground and background, separating overlapping objects, classifying objects, and performing quantitative calculations to generate accurate object information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If objects in contact with each other are processed as a single entity, then processing complexity is reduced, but measurement precision of individual objects deteriorates
Solution Approach 1:
The patent applies segmentation by dividing the foreground region into multiple independent object regions. The separation unit segments the foreground according to contact relationships between objects, creating distinct regions for each object even when they are in contact. This allows individual objects to be identified and measured separately, resolving the contradiction between processing simplicity and measurement accuracy.
2Productivity
If simple object counting methods are used, then processing speed is improved, but reliability of object classification deteriorates
Solution Approach 1:
The patent applies preliminary action by performing foreground-background segmentation and object separation before classification and counting. The segmentation unit and separation unit prepare the image data in advance by identifying contact relationships and creating separate object regions. This preliminary processing enables reliable classification and counting in subsequent steps without significantly increasing overall processing time.
Data Source
AI summary
In order to perform quantitative analysis on an object in an image, it is important to accurately identify the object, but when plural objects are in contact with each other, it is potential that a target portion cannot be accurately identified. An image is segmented into a foreground region and a background region, the foreground region being a region in which an object for which quantitative information is to be calculated is shown, and the background region being a region other than the foreground region. With respect to a first object and a second object in contact with each other in the image, a contact point between the first object and the second object is detected based on a region segmentation result output by a segmentation unit. The first object and the second object can be separated by connecting two boundary reference pixels including a first boundary reference pixel that is a pixel in a background region closest to the contact point, and a second boundary reference pixel that is a pixel in a background region in a direction opposite to the first boundary reference pixel across the contact point.