Endoscope Image Processing for Shape-Based Color Assignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing image processing techniques for endoscope images captured with white light have limited color assignment options, making it difficult to effectively differentiate and visualize the shape features of structures like blood vessels and gland ducts in the mucosa, which is crucial for medical diagnosis.
Innovation Solution
An image processing apparatus and method that includes an image signal acquisition unit, an original image generation unit, a reduced original image generation unit, a structure region image generation unit, a feature amount calculation unit, a color assignment unit, and an image combining unit, which allows for high-degree freedom of color assignment based on shape features by extracting structure regions and assigning colors corresponding to calculated feature amounts, enabling the generation of composite images that emphasize these structures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a pseudo-color image is generated by assigning colors to a normal color image captured with white light, then structure regions such as blood vessels can be emphasized, but the number of assignable colors is limited and the degree of freedom of color assignment is low
Solution Approach 1:
The patent transitions from 2D color assignment (RGB channels) to 3D color space utilization (L*a*b* space with L*, a*, b* dimensions). This dimensional expansion allows multiple feature amounts to be mapped to different color dimensions simultaneously, enabling high-degree freedom color assignment while maintaining diagnostic reliability through systematic color-feature correspondence.
Solution Approach 2:
The patent changes the color representation parameters from standard RGB to L*a*b* color space, and further transforms color values based on multiple feature amounts (brightness, saturation, hue adjustments). This parameter transformation enables continuous color variation according to different shape features, resolving the limitation of fixed color assignment in conventional methods.
2Loss of information
If multiple structure regions with different shape features are extracted and classified, then detailed visualization of structures like blood vessels and gland ducts is achieved, but the complexity of image processing increases
Solution Approach 1:
The patent segments the image processing into distinct functional modules: structure region extraction unit, feature amount calculation unit, color assignment unit, and composite image generation unit. Each module handles a specific aspect of processing, which reduces overall system complexity by localizing functions while maintaining comprehensive shape feature analysis.
Solution Approach 2:
The patent introduces a feature amount calculation unit as an intermediary between structure region extraction and color assignment. This intermediary computes multiple shape features (area, perimeter, circularity, etc.) and uses them as parameters for systematic color assignment, simplifying the connection between complex extraction results and visual representation.
3Stability of the object's composition
If the original color image is used as the base for color assignment, then natural colors are preserved, but the number of colors that can be assigned is restricted by the existing color palette
Solution Approach 1:
The patent transforms the color parameters of the original image by adjusting brightness (L*), saturation (a*), and hue (b*) based on multiple feature amounts. This parameter modification enables the generation of differentiated colors while maintaining the structural consistency of the original image, allowing up to 3D color variation without completely replacing the original color information.
Solution Approach 2:
The patent makes the color assignment system universal by enabling a single processing framework to handle multiple structure types (blood vessels, gland ducts, etc.) and multiple shape features simultaneously. The L*a*b* color space serves as a universal canvas that can represent diverse color assignments while maintaining compatibility with the original image structure.
Data Source
AI summary
There is provided an image processing apparatus, method, and program for an endoscope image having a high degree of freedom of color assignment according to the shape feature of a structure. A structure region image is generated by extracting one or more structure regions from an image signal obtained by imaging an observation target in the living body. A shape feature amount of the structure region is calculated based on the structure region image, and a color corresponding to the feature amount is assigned to the structure region. A reduced original image is generated by performing processing for reducing at least one of the color or the contrast on an original image in which the observation target is drawn in color. A composite image is generated by superimposing the reduced original image and the structure region image subjected to the color assignment.