Dermoscopic Image Processing for Lesion Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current dermoscopic diagnosis methods face challenges in achieving accurate discrimination between benign and malignant tumors, relying heavily on physician skill and lacking consistent diagnostic accuracy due to uniform image processing regardless of lesion type or progression stage.
Innovation Solution
A diagnostic apparatus and method that utilizes a dermoscope-equipped image-capturing device connected to a processing unit for image classification, employing color component analysis, structure clarifying, region highlighting, and region fluorescent coloring conversions to enhance diagnostic accuracy by providing processed images that highlight specific features of skin lesions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If uniform image conversion is performed on all dermoscopic images regardless of lesion type or stage, then the processing is simple and fast, but diagnostic accuracy is insufficient and relies heavily on physician skill
Solution Approach 1:
The patent segments the image processing into multiple stages: initial image acquisition, classification based on lesion type and progression stage, and then application of specific conversion processes appropriate to each class. This segmentation allows complex processing to be applied selectively rather than uniformly, improving diagnostic accuracy while managing complexity through structured organization.
Solution Approach 2:
The patent performs preliminary classification of the dermoscopic image to determine lesion type and progression stage before applying the appropriate conversion process. This preliminary action enables the system to select the most suitable processing method in advance, ensuring optimal enhancement for each specific case rather than using a one-size-fits-all approach.
Solution Approach 3:
The patent changes processing parameters based on the classified lesion characteristics. Different conversion processes with distinct parameters are applied according to the lesion type and stage, allowing the system to optimize image enhancement parameters for each specific diagnostic scenario, thereby improving measurement precision without requiring all possible complex processing to be available simultaneously.
2Measurement precision
If multiple specialized image conversion processes are developed for different lesion types and stages, then diagnostic accuracy improves, but the device complexity and processing time increase
Solution Approach 1:
The classification step is performed as a preliminary action that quickly categorizes the lesion type and progression stage before applying specialized conversion processes. This preliminary classification enables the system to select the appropriate processing path efficiently, avoiding unnecessary processing steps and minimizing overall processing time while maintaining high diagnostic accuracy through targeted enhancement.
Solution Approach 2:
The patent segments the processing into distinct stages where each stage handles a specific aspect: classification, then selective conversion application. This segmentation allows the system to apply only the necessary specialized processes for each lesion type and stage, avoiding redundant processing and reducing overall time loss while maintaining diagnostic precision.
3Loss of information
If dermoscopic inspection with magnification is used to eliminate scattered reflection and visualize pigmentation distribution, then visibility of skin layer structures improves, but discrimination between benign and malignant tumors remains difficult
Solution Approach 1:
The patent applies parameter changes through specialized image conversion processes that enhance specific features relevant to tumor discrimination. By adjusting enhancement parameters according to lesion type and progression stage, the system optimizes the visibility of discriminative features such as pigmentation patterns, border characteristics, and structural organization, thereby improving discrimination accuracy beyond what standard magnification alone can achieve.
Solution Approach 2:
The patent applies different conversion processes with different enhancement characteristics to different types of lesions based on their specific features. This local quality approach ensures that each lesion type receives the most appropriate enhancement tailored to its characteristics, maximizing the visibility of discriminative features for that specific lesion type rather than using a uniform enhancement approach.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention provides a method of processing an image in a diagnostic apparatus 100 of diagnosing a lesion using a captured image of an affected area to be diagnosed, comprising the steps of: (i) classifying the captured image based on a stage of a progression of the lesion (Step S13 and S14); and (ii) performing an image conversion processing, which corresponds to a classification obtained as a result of step (i), on the captured image to generate a converted image (Step S15-17).