Skin Lesion Tracking Across Imaging Modalities and Time
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for tracking skin lesions are inefficient, and existing methods fail to accurately detect and classify skin lesions due to the lack of automation in dermatology, resulting in inefficiencies and errors in dermatology. Existing methods fail to effectively detect and classify skin lesions due to the lack of automation in dermatology.
Innovation Solution
A method and apparatus for tracking changes in skin features over time, across different imaging modalities and devices, using hierarchical motion estimation and spatio-temporal analysis to automate the identification and tracking of skin lesions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual annotation and tracking of skin lesions is performed by dermatologists and their staff, then diagnostic accuracy can be maintained through expert evaluation, but the process becomes extremely time-consuming and labor-intensive
Solution Approach 1:
The patent replaces the manual mechanical process of dermatologists visually examining and annotating skin lesions with an automated computer vision system. The system uses image processing algorithms to automatically detect, segment, and track skin lesions across multiple images, eliminating the need for manual annotation while maintaining diagnostic accuracy through sophisticated pattern recognition and change detection capabilities.
Solution Approach 2:
The system enables self-service by allowing the computer vision algorithm to autonomously perform lesion detection, tracking, and change detection without requiring continuous human intervention. The automated system processes images, identifies lesions, and generates diagnostic reports independently, freeing dermatologists from tedious manual tasks while preserving expert oversight for final diagnosis.
2Loss of information
If multiple imaging modalities and devices are used to capture skin images over time, then comprehensive diagnostic information can be obtained, but the complexity of processing and comparing these images increases significantly
Solution Approach 1:
The patent implements a universal image processing framework that can handle multiple imaging modalities (dermoscopy, clinical photography, microscopy) and various devices within a single system. The standardized processing pipeline normalizes images from different sources, enabling comprehensive diagnostic information to be integrated and compared across timepoints without requiring separate processing workflows for each modality or device type.
Solution Approach 2:
The system dynamically adjusts processing parameters based on the specific imaging modality and device used. By automatically detecting image characteristics and adapting processing algorithms accordingly, the system maintains optimal performance across diverse imaging conditions while simplifying the overall processing complexity through intelligent parameter adaptation rather than requiring manual configuration for each device type.
3Reliability
If manual tracking of skin lesions across multiple patient visits is performed, then lesion evolution can be monitored, but the process becomes overwhelming and error-prone
Solution Approach 1:
The patent replaces the manual tracking process with an automated system that uses image registration and feature matching algorithms to automatically monitor lesion evolution across multiple patient visits. The system consistently identifies and tracks lesions through time, eliminating human errors in manual tracking while maintaining reliable monitoring of lesion changes, and presents results in an intuitive format that is easy to interpret.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Methods and apparatus are disclosed that assist a user such as a doctor to track changes that occur in features of a subject's skin as the skin features evolve over time. Such a tracking can be done for images captured under different imaging/lighting modalities, by different image capture devices, and/or at different points in time. Methods and apparatus to automatically identify and track the unconstrained appearance/disappearance of skin features are disclosed.